retell-sdk 6.0.1 → 6.1.0

package/resources/chat-agent.d.ts

@@ -134,6 +134,14 @@ export interface ChatAgentResponse {
      * Version that this draft was based on. Null for initial versions.
      */
     base_version?: number | null;
+    /**
+     * Contact memory settings for phone calls and SMS chats. Creating an agent
+     * defaults enable_update to false and enable_read to true. Updates only change the
+     * supplied flags; omitted flags stay unchanged and an empty object has no effect.
+     * Set a flag to false to disable it. The configuration cannot be cleared. Existing
+     * agents without this configuration have both disabled.
+     */
+    contact_memory_config?: ChatAgentResponse.ContactMemoryConfig;
     /**
      * Number of days to retain call/chat data before automatic deletion. Must be
      * between 1 and 730 days. If not set, data is retained forever (no automatic
@@ -194,7 +202,17 @@ export interface ChatAgentResponse {
     /**
      * The model to use for post chat analysis. Default to gpt-5.6-terra.
      */
-    post_chat_analysis_model?: 'gpt-4.1' | 'gpt-4.1-mini' | 'gpt-4.1-nano' | 'gpt-5' | 'gpt-5-mini' | 'gpt-5-nano' | 'gpt-5.1' | 'gpt-5.2' | 'gpt-5.4' | 'gpt-5.4-mini' | 'gpt-5.4-nano' | 'gpt-5.5' | 'gpt-5.6-terra' | 'gpt-5.6-luna' | 'claude-4.5-sonnet' | 'claude-4.6-sonnet' | 'claude-5-sonnet' | 'claude-4.5-haiku' | 'gemini-3.0-flash' | 'gemini-3.1-flash-lite' | 'gemini-3.5-flash' | 'gemini-3.5-flash-lite' | 'gemini-3.6-flash' | 'gemini-3.7-flash' | 'gemini-3.8-flash' | null;
+    post_chat_analysis_model?: 'gpt-4.1' | 'gpt-4.1-mini' | 'gpt-4.1-nano' | 'gpt-5' | 'gpt-5-mini' | 'gpt-5-nano' | 'gpt-5.1' | 'gpt-5.2' | 'gpt-5.4' | 'gpt-5.4-mini' | 'gpt-5.4-nano' | 'gpt-5.5' | 'gpt-5.6-terra' | 'gpt-5.6-luna' | 'gpt-6-astra' | 'gpt-6-sol' | 'gpt-6.1-sol' | 'gpt-6-luna' | 'claude-4.5-sonnet' | 'claude-4.6-sonnet' | 'claude-5-opus' | 'claude-5.5-opus' | 'claude-5-sonnet' | 'claude-5.5-sonnet' | 'claude-4.5-haiku' | 'gemini-3.0-flash' | 'gemini-3.1-flash-lite' | 'gemini-3.5-flash' | 'gemini-3.5-flash-lite' | 'gemini-3.6-flash' | 'gemini-3.7-flash' | 'gemini-3.8-flash' | null;
+    /**
+     * Integration (Agent Functions) tools run as a dependency graph at chat end, after
+     * post-chat analysis. Each tool can be gated by a condition. Set to null to clear.
+     */
+    post_session_tools?: Array<ChatAgentResponse.AppTool | ChatAgentResponse.CustomTool | ChatAgentResponse.CodeTool | ChatAgentResponse.SendSMSTool> | null;
+    /**
+     * Integration (Agent Functions) tools run as a dependency graph before the chat's
+     * first message. Outputs are injected as dynamic variables. Set to null to clear.
+     */
+    pre_session_tools?: Array<ChatAgentResponse.AppTool | ChatAgentResponse.CustomTool | ChatAgentResponse.CodeTool> | null;
     /**
      * The expiration time for the signed url in milliseconds. Only applicable when
      * opt_in_signed_url is true. If not set, default value of 86400000 (24 hours) will
@@ -271,6 +289,27 @@ export declare namespace ChatAgentResponse {
          */
         version?: number | null;
     }
+    /**
+     * Contact memory settings for phone calls and SMS chats. Creating an agent
+     * defaults enable_update to false and enable_read to true. Updates only change the
+     * supplied flags; omitted flags stay unchanged and an empty object has no effect.
+     * Set a flag to false to disable it. The configuration cannot be cleared. Existing
+     * agents without this configuration have both disabled.
+     */
+    interface ContactMemoryConfig {
+        /**
+         * Automatically add saved contact memory to the agent prompt. Skippable nodes can
+         * use answers from the current conversation even when this setting is disabled.
+         * Contact dynamic variables, including contact_memory, remain available regardless
+         * of this setting.
+         */
+        enable_read?: boolean;
+        /**
+         * Rewrite the contact memory after each conversation. Requires storing
+         * conversation data. Chat agents must also have end_chat_after_silence_ms set.
+         */
+        enable_update?: boolean;
+    }
     /**
      * Configuration for guardrail checks to detect and prevent prohibited topics in
      * agent output and user input.
@@ -461,281 +500,2374 @@ export declare namespace ChatAgentResponse {
          */
         required?: boolean;
     }
-}
-export interface ChatAgentListResponse {
-    /**
-     * Whether more results are available.
-     */
-    has_more: boolean;
-    items: Array<ChatAgentListResponse.Item>;
-    /**
-     * Pagination key for the next page.
-     */
-    pagination_key?: string;
-}
-export declare namespace ChatAgentListResponse {
-    interface Item {
+    interface AppTool {
         /**
-         * Unique id of agent.
+         * The connection (App) this tool runs against. Must be a connection in the
+         * organization whose provider matches this tool's provider.
          */
-        agent_id: string;
+        app_id: string;
         /**
-         * The name of the agent. Only used for your own reference.
+         * Name of the catalog template within the provider, as listed by
+         * list-app-templates.
          */
-        agent_name: string;
-        channel: 'voice' | 'chat';
+        app_tool_template_name: string;
         /**
-         * Authoritative root tags for this agent, keyed by tag name.
+         * Name of the tool. Must be unique within the phase's tools; referenced by
+         * depends_on.
          */
-        tags: {
-            [key: string]: Item.Tags;
-        };
+        name: string;
         /**
-         * User modification timestamp (milliseconds since epoch). Either the time of last
-         * update or creation if no updates available.
+         * Provider of the connection. Must match the connection's provider; supported
+         * providers are listed by list-app-templates.
          */
-        user_modified_timestamp: number;
-    }
-    namespace Item {
-        interface Tags {
-            dynamic_variables?: {
-                [key: string]: string;
-            };
-            version?: number;
-        }
-    }
-}
-export type ChatAgentCreateVersionResponse = AgentAPI.AgentResponse | ChatAgentResponse;
-export interface ChatAgentCreateParams {
-    /**
-     * The Response Engine to attach to the agent. It is used to generate responses for
-     * the agent. You need to create a Response Engine first before attaching it to an
-     * agent.
-     */
-    response_engine: ChatAgentCreateParams.ResponseEngineRetellLm | ChatAgentCreateParams.ResponseEngineCustomLm | ChatAgentCreateParams.ResponseEngineConversationFlow;
-    /**
-     * The name of the chat agent. Only used for your own reference.
-     */
-    agent_name?: string | null;
-    /**
-     * Message to display when the chat is automatically closed.
-     */
-    auto_close_message?: string | null;
-    /**
-     * Number of days to retain call/chat data before automatic deletion. Must be
-     * between 1 and 730 days. If not set, data is retained forever (no automatic
-     * deletion).
-     */
-    data_storage_retention_days?: number | null;
-    /**
-     * Controls what data is stored for this agent. "everything" stores all data
-     * including transcripts and recordings. "everything_except_pii" stores data but
-     * excludes PII when possible based on PII configuration. "basic_attributes_only"
-     * stores only basic metadata. If not set, defaults to "everything".
-     */
-    data_storage_setting?: 'everything' | 'everything_except_pii' | 'basic_attributes_only' | null;
-    /**
-     * If users stay silent for a period after agent speech, end the chat. The minimum
-     * value allowed is 120,000 ms (2 minutes). The maximum value allowed is
-     * 259,200,000 ms (72 hours). By default, this is set to 3,600,000 (1 hour).
-     */
-    end_chat_after_silence_ms?: number | null;
-    /**
-     * Configuration for guardrail checks to detect and prevent prohibited topics in
-     * agent output and user input.
-     */
-    guardrail_config?: ChatAgentCreateParams.GuardrailConfig;
-    /**
-     * Toggle behavior presets on/off to influence agent response style and behaviors.
-     * Voice-only presets are not available for chat agents.
-     */
-    handbook_config?: ChatAgentCreateParams.HandbookConfig;
-    /**
-     * Specifies what language(s) the agent will operate in. Accepts either a single
-     * locale (e.g. `en-US`) or an array of locales for multilingual agents (e.g.
-     * `["en-US","es-ES"]`). The scalar value `multi` is deprecated but still accepted
-     * as a scalar, and is stored and returned as the ten locales it used to mean. It
-     * must not appear inside the array form. Send an explicit locale array instead. If
-     * unset, defaults to `en-US`.
-     */
-    language?: 'en-US' | 'en-IN' | 'en-GB' | 'en-AU' | 'en-NZ' | 'de-DE' | 'es-ES' | 'es-419' | 'hi-IN' | 'fr-FR' | 'fr-CA' | 'ja-JP' | 'pt-PT' | 'pt-BR' | 'zh-CN' | 'ru-RU' | 'it-IT' | 'ko-KR' | 'nl-NL' | 'nl-BE' | 'pl-PL' | 'tr-TR' | 'vi-VN' | 'ro-RO' | 'bg-BG' | 'ca-ES' | 'th-TH' | 'da-DK' | 'fi-FI' | 'el-GR' | 'hu-HU' | 'id-ID' | 'no-NO' | 'sk-SK' | 'sv-SE' | 'lt-LT' | 'lv-LV' | 'cs-CZ' | 'ms-MY' | 'af-ZA' | 'ar-SA' | 'az-AZ' | 'bs-BA' | 'cy-GB' | 'fa-IR' | 'fil-PH' | 'gl-ES' | 'he-IL' | 'hr-HR' | 'hy-AM' | 'is-IS' | 'kk-KZ' | 'kn-IN' | 'mk-MK' | 'mr-IN' | 'ne-NP' | 'sl-SI' | 'sr-RS' | 'sw-KE' | 'ta-IN' | 'ur-IN' | 'yue-CN' | 'uk-UA' | 'multi' | Array<'en-US' | 'en-IN' | 'en-GB' | 'en-AU' | 'en-NZ' | 'de-DE' | 'es-ES' | 'es-419' | 'hi-IN' | 'fr-FR' | 'fr-CA' | 'ja-JP' | 'pt-PT' | 'pt-BR' | 'zh-CN' | 'ru-RU' | 'it-IT' | 'ko-KR' | 'nl-NL' | 'nl-BE' | 'pl-PL' | 'tr-TR' | 'vi-VN' | 'ro-RO' | 'bg-BG' | 'ca-ES' | 'th-TH' | 'da-DK' | 'fi-FI' | 'el-GR' | 'hu-HU' | 'id-ID' | 'no-NO' | 'sk-SK' | 'sv-SE' | 'lt-LT' | 'lv-LV' | 'cs-CZ' | 'ms-MY' | 'af-ZA' | 'ar-SA' | 'az-AZ' | 'bs-BA' | 'cy-GB' | 'fa-IR' | 'fil-PH' | 'gl-ES' | 'he-IL' | 'hr-HR' | 'hy-AM' | 'is-IS' | 'kk-KZ' | 'kn-IN' | 'mk-MK' | 'mr-IN' | 'ne-NP' | 'sl-SI' | 'sr-RS' | 'sw-KE' | 'ta-IN' | 'ur-IN' | 'yue-CN' | 'uk-UA'>;
-    /**
-     * Whether this agent opts in to signed url for public log. If not set, default
-     * value of false will apply.
-     */
-    opt_in_signed_url?: boolean;
-    /**
-     * Configuration for PII scrubbing from transcripts and recordings.
-     */
-    pii_config?: ChatAgentCreateParams.PiiConfig;
-    /**
-     * Post chat analysis data to extract from the chat. This data will augment the
-     * pre-defined variables extracted in the chat analysis. This will be available
-     * after the chat ends.
-     */
-    post_chat_analysis_data?: Array<ChatAgentCreateParams.StringAnalysisData | ChatAgentCreateParams.EnumAnalysisData | ChatAgentCreateParams.BooleanAnalysisData | ChatAgentCreateParams.NumberAnalysisData | ChatAgentCreateParams.ChatPresetAnalysisData> | null;
-    /**
-     * The model to use for post chat analysis. Default to gpt-5.6-terra.
-     */
-    post_chat_analysis_model?: 'gpt-4.1' | 'gpt-4.1-mini' | 'gpt-4.1-nano' | 'gpt-5' | 'gpt-5-mini' | 'gpt-5-nano' | 'gpt-5.1' | 'gpt-5.2' | 'gpt-5.4' | 'gpt-5.4-mini' | 'gpt-5.4-nano' | 'gpt-5.5' | 'gpt-5.6-terra' | 'gpt-5.6-luna' | 'claude-4.5-sonnet' | 'claude-4.6-sonnet' | 'claude-5-sonnet' | 'claude-4.5-haiku' | 'gemini-3.0-flash' | 'gemini-3.1-flash-lite' | 'gemini-3.5-flash' | 'gemini-3.5-flash-lite' | 'gemini-3.6-flash' | 'gemini-3.7-flash' | 'gemini-3.8-flash' | null;
-    /**
-     * The expiration time for the signed url in milliseconds. Only applicable when
-     * opt_in_signed_url is true. If not set, default value of 86400000 (24 hours) will
-     * apply.
-     */
-    signed_url_expiration_ms?: number | null;
-    /**
-     * IANA timezone for the agent (e.g. America/New_York). Defaults to
-     * America/Los_Angeles if not set.
-     */
-    timezone?: string | null;
-    /**
-     * Optional title of the chat agent version. Used for your own reference.
-     */
-    version_title?: string | null;
-    /**
-     * Which webhook events this agent should receive. If not set, defaults to
-     * chat_started, chat_ended, chat_analyzed.
-     */
-    webhook_events?: Array<'chat_started' | 'chat_ended' | 'chat_analyzed' | 'transcript_updated'> | null;
-    /**
-     * The timeout for the webhook in milliseconds. If not set, default value of 10000
-     * will apply.
-     */
-    webhook_timeout_ms?: number;
-    /**
-     * The webhook for agent to listen to chat events. See what events it would get at
-     * [webhook doc](/features/webhook). If set, will binds webhook events for this
-     * agent to the specified url, and will ignore the account level webhook for this
-     * agent. Set to `null` to remove webhook url from this agent.
-     */
-    webhook_url?: string | null;
-}
-export declare namespace ChatAgentCreateParams {
-    interface ResponseEngineRetellLm {
+        provider: string;
+        type: 'integration_app';
         /**
-         * id of the Retell LLM Response Engine.
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
          */
-        llm_id: string;
+        condition?: AppTool.Condition;
         /**
-         * type of the Response Engine.
+         * Names of tools that must run before this one.
          */
-        type: 'retell-llm';
+        depends_on?: Array<string>;
         /**
-         * Version of the Retell LLM Response Engine.
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Overrides the catalog template's LLM-facing description.
          */
-        version?: number | null;
-    }
-    interface ResponseEngineCustomLm {
+        description?: string;
         /**
-         * LLM websocket url of the custom LLM.
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. If true, play a typing sound on the agent audio track while this
+         * tool is executing. Useful when the tool takes a noticeable amount of time to
+         * prevent silence on the call.
          */
-        llm_websocket_url: string;
+        enable_typing_sound?: boolean;
         /**
-         * type of the Response Engine.
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. The message for the agent to speak when executing the tool. Only
+         * applicable when speak_during_execution is true.
          */
-        type: 'custom-llm';
-    }
-    interface ResponseEngineConversationFlow {
+        execution_message_description?: string;
         /**
-         * ID of the Conversation Flow Response Engine.
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
          */
-        conversation_flow_id: string;
+        execution_message_type?: 'prompt' | 'static_text';
         /**
-         * type of the Response Engine.
+         * What the agent and the transcript see of the tool's response. Omit to send the
+         * full response. Does not affect response_variables, which are always extracted
+         * from the raw response.
          */
-        type: 'conversation-flow';
+        output_selection?: AppTool.UnionMember0 | AppTool.UnionMember1;
         /**
-         * Version of the Conversation Flow Response Engine.
+         * The resolved input parameters, in order. Properties may pin a value with const
+         * (including {{variable}} references) or provide a description for LLM inference.
+         * Each property may also record selected*input_mode, the editor mode the user
+         * selected ("const_enum", "const_boolean", "const_value", "description_custom", or
+         * "description_preset"); it is stored and returned as-is, used only by the tool
+         * config UI. Omit the key when no mode is recorded; when set, const*_ modes
+         * require a non-empty const, and description\__ modes must omit const entirely.
+         * Each parameter's required list must match the schema returned by the
+         * corresponding step of the get-app-tool-schema loop.
          */
-        version?: number | null;
-    }
-    /**
-     * Configuration for guardrail checks to detect and prevent prohibited topics in
-     * agent output and user input.
-     */
-    interface GuardrailConfig {
+        parameters?: Array<AppTool.Parameter>;
         /**
-         * Selected prohibited user topic categories to check. When user messages contain
-         * these topics, the agent will respond with a placeholder message instead of
-         * processing the request.
+         * Mapping of a dynamic-variable name to the response field (dot-path) it is
+         * populated from. Missing paths are ignored.
          */
-        input_topics?: Array<'platform_integrity_jailbreaking'> | null;
+        response_variables?: {
+            [key: string]: string;
+        };
         /**
-         * Selected prohibited agent topic categories to check. When agent messages contain
-         * these topics, they will be replaced with a placeholder message.
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Determines whether the agent would call LLM another time and speak
+         * when the result of the tool is obtained.
          */
-        output_topics?: Array<'harassment' | 'self_harm' | 'sexual_exploitation' | 'violence' | 'defense_and_national_security' | 'illicit_and_harmful_activity' | 'gambling' | 'regulated_professional_advice' | 'child_safety_and_exploitation'> | null;
-    }
-    /**
-     * Toggle behavior presets on/off to influence agent response style and behaviors.
-     * Voice-only presets are not available for chat agents.
-     */
-    interface HandbookConfig {
+        speak_after_execution?: boolean;
         /**
-         * When asked, acknowledge being a virtual assistant.
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. If true, will speak during execution.
          */
-        ai_disclosure?: boolean;
+        speak_during_execution?: boolean;
+    }
+    namespace AppTool {
         /**
-         * Professional call center rep baseline.
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
          */
-        default_personality?: boolean;
+        interface Condition {
+            equations: Array<Condition.Equation>;
+            operator: '||' | '&&';
+            type: 'equation';
+        }
+        namespace Condition {
+            interface Equation {
+                /**
+                 * Left side of the equation
+                 */
+                left: string;
+                operator: '==' | '!=' | '>' | '>=' | '<' | '<=' | 'contains' | 'not_contains' | 'exists' | 'not_exist';
+                /**
+                 * Right side of the equation. The right side of the equation not required when
+                 * "exists" or "not_exist" are selected.
+                 */
+                right?: string;
+            }
+        }
+        interface UnionMember0 {
+            mode: 'all';
+            /**
+             * Not used at runtime; stored and returned as-is for the UI.
+             */
+            fields?: Array<string>;
+        }
+        interface UnionMember1 {
+            /**
+             * The only response fields the agent and the transcript see, as dot-paths into the
+             * response schema returned by get-app-tool-schema. Everything else is dropped.
+             * Selecting a parent keeps its whole subtree. A plain segment traverses arrays
+             * element-wise (deals.properties.amount keeps that field on every deal), while
+             * key[n] selects one element (deals[0].id keeps only the first deal's id); paths
+             * that match nothing contribute nothing.
+             */
+            fields: Array<string>;
+            mode: 'subset';
+        }
         /**
-         * Warm acknowledgment of caller concerns.
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
          */
-        high_empathy?: boolean;
+        interface Parameter {
+            /**
+             * The value of properties is an object, where each key is the name of a property
+             * and each value is a schema used to validate that property.
+             */
+            properties: unknown;
+            /**
+             * Type must be "object" for a JSON Schema object.
+             */
+            type: 'object';
+            /**
+             * List of names of required property when generating this parameter. LLM will do
+             * its best to generate the required properties in its function arguments. Property
+             * must exist in properties.
+             */
+            required?: Array<string>;
+        }
+    }
+    interface CustomTool {
         /**
-         * Stay within prompt/context scope, don't invent details.
+         * Name of the tool. Must be unique within all tools available to LLM at any given
+         * time (general tools + state tools + state edges). Must be consisted of a-z, A-Z,
+         * 0-9, or contain underscores and dashes, with a maximum length of 64 (no space
+         * allowed).
          */
-        scope_boundaries?: boolean;
-    }
-    /**
-     * Configuration for PII scrubbing from transcripts and recordings.
-     */
-    interface PiiConfig {
+        name: string;
+        type: 'custom';
         /**
-         * List of PII categories to scrub from transcripts and recordings. PII redaction
-         * is only active when this list is non-empty; an empty array means no PII
-         * scrubbing is performed.
+         * Describes what the tool does, sometimes can also include information about when
+         * to call the tool.
          */
-        categories: Array<'person_name' | 'address' | 'email' | 'phone_number' | 'ssn' | 'passport' | 'driver_license' | 'credit_card' | 'bank_account' | 'password' | 'pin' | 'medical_id' | 'date_of_birth' | 'customer_account_number'>;
+        url: string;
         /**
-         * The processing mode for PII scrubbing. Currently only post-call is supported.
+         * If set to true, the parameters will be passed as root level JSON object instead
+         * of nested under "args".
          */
-        mode: 'post_call';
-    }
-    interface StringAnalysisData {
+        args_at_root?: boolean;
         /**
-         * Description of the variable.
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
          */
-        description: string;
+        condition?: CustomTool.Condition;
         /**
-         * Name of the variable.
+         * Names of tools that must run before this one.
          */
-        name: string;
+        depends_on?: Array<string>;
         /**
-         * Type of the variable to extract.
+         * Describes what this tool does and when to call this tool.
          */
-        type: 'string';
+        description?: string;
         /**
-         * Optional instruction to help decide whether this field needs to be populated in
-         * the analysis. If not set, the field is always included. If required is true,
-         * this is ignored.
+         * If true, play a typing sound on the agent audio track while this tool is
+         * executing. Useful when the tool takes a noticeable amount of time to prevent
+         * silence on the call.
          */
-        conditional_prompt?: string;
+        enable_typing_sound?: boolean;
         /**
-         * Examples of the variable value to teach model the style and syntax.
+         * The description for the sentence agent say during execution. Only applicable
+         * when speak_during_execution is true. Can write what to say or even provide
+         * examples. The default is "The message you will say to callee when calling this
+         * tool. Make sure it fits into the conversation smoothly.".
          */
-        examples?: Array<string>;
+        execution_message_description?: string;
         /**
-         * Whether this data is required. If true and the data is not extracted, the call
-         * will be marked as unsuccessful.
+         * Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
          */
-        required?: boolean;
-    }
-    interface EnumAnalysisData {
+        execution_message_type?: 'prompt' | 'static_text';
         /**
-         * The possible values of the variable, must be non empty array.
+         * Headers to add to the request.
+         */
+        headers?: {
+            [key: string]: string;
+        };
+        /**
+         * Maximum number of times to retry the request after a failed attempt, from 0 (no
+         * retry) to 5. Retries happen on any failure, with exponential backoff between
+         * attempts; the backoff delay is not configurable. `timeout_ms` applies per
+         * attempt rather than as a budget across all attempts, so an attempt that times
+         * out is still retried and the worst-case total duration is `timeout_ms`
+         * multiplied by (`max_retry` + 1) as well as any latency incurred by the
+         * exponential backoff + jitter between each retry. Only the final attempt's result
+         * is reported to the agent. Because retries repeat the request, only set this
+         * above 0 if your endpoint is idempotent — a retried request may be processed more
+         * than once. Defaults to 0 (no retry).
+         */
+        max_retry?: number;
+        /**
+         * Method to use for the request, default to POST.
+         */
+        method?: 'GET' | 'POST' | 'PUT' | 'PATCH' | 'DELETE';
+        /**
+         * How the tool's `parameters` are authored and shown in the dashboard editor —
+         * "form" for the visual parameter builder, "json" for a raw JSON Schema. Both
+         * produce the same `parameters` schema; this does not change how the request body
+         * is encoded (see `args_at_root`).
+         */
+        parameter_type?: 'json' | 'form';
+        /**
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
+         */
+        parameters?: CustomTool.Parameters;
+        /**
+         * Query parameters to append to the request URL.
+         */
+        query_params?: {
+            [key: string]: string;
+        };
+        /**
+         * A mapping of variable names to JSON paths in the response body. These values
+         * will be extracted from the response and made available as dynamic variables for
+         * use.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Determines whether the agent would call LLM another time and speak when the
+         * result of function is obtained. Usually this needs to get turned on so user can
+         * get update for the function call.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Determines whether the agent would say sentence like "One moment, let me check
+         * that." when executing the function. Recommend to turn on if your function call
+         * takes over 1s (including network) to complete, so that your agent remains
+         * responsive.
+         */
+        speak_during_execution?: boolean;
+        /**
+         * The maximum time in milliseconds the tool can run before it's considered
+         * timeout. If the tool times out, the agent would have that info. The minimum
+         * value allowed is 1000 ms (1 s), and maximum value allowed is 600,000 ms (10
+         * min). By default, this is set to 120,000 ms (2 min).
+         */
+        timeout_ms?: number;
+    }
+    namespace CustomTool {
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        interface Condition {
+            equations: Array<Condition.Equation>;
+            operator: '||' | '&&';
+            type: 'equation';
+        }
+        namespace Condition {
+            interface Equation {
+                /**
+                 * Left side of the equation
+                 */
+                left: string;
+                operator: '==' | '!=' | '>' | '>=' | '<' | '<=' | 'contains' | 'not_contains' | 'exists' | 'not_exist';
+                /**
+                 * Right side of the equation. The right side of the equation not required when
+                 * "exists" or "not_exist" are selected.
+                 */
+                right?: string;
+            }
+        }
+        /**
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
+         */
+        interface Parameters {
+            /**
+             * The value of properties is an object, where each key is the name of a property
+             * and each value is a schema used to validate that property.
+             */
+            properties: unknown;
+            /**
+             * Type must be "object" for a JSON Schema object.
+             */
+            type: 'object';
+            /**
+             * List of names of required property when generating this parameter. LLM will do
+             * its best to generate the required properties in its function arguments. Property
+             * must exist in properties.
+             */
+            required?: Array<string>;
+        }
+    }
+    interface CodeTool {
+        /**
+         * JavaScript code to execute in the sandbox.
+         */
+        code: string;
+        /**
+         * Name of the tool. Must be unique within all tools available to LLM at any given
+         * time (general tools + state tools + state edges). Must be consisted of a-z, A-Z,
+         * 0-9, or contain underscores and dashes, with a maximum length of 64 (no space
+         * allowed).
+         */
+        name: string;
+        type: 'code';
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        condition?: CodeTool.Condition;
+        /**
+         * Names of tools that must run before this one.
+         */
+        depends_on?: Array<string>;
+        /**
+         * Describes what this tool does and when to call this tool.
+         */
+        description?: string;
+        /**
+         * If true, play a typing sound on the agent audio track while this tool is
+         * executing.
+         */
+        enable_typing_sound?: boolean;
+        /**
+         * The description for the sentence agent say during execution. Only applicable
+         * when speak_during_execution is true.
+         */
+        execution_message_description?: string;
+        /**
+         * Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * A mapping of variable names to JSON paths in the code execution result. These
+         * mapped values will be extracted and added as dynamic variables.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Determines whether the agent would call LLM another time and speak when the
+         * result of function is obtained.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Determines whether the agent would say sentence like "One moment, let me check
+         * that." when executing the tool.
+         */
+        speak_during_execution?: boolean;
+        /**
+         * The maximum time in milliseconds the code can run before it's considered
+         * timeout. Defaults to 30,000 ms (30 s).
+         */
+        timeout_ms?: number;
+    }
+    namespace CodeTool {
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        interface Condition {
+            equations: Array<Condition.Equation>;
+            operator: '||' | '&&';
+            type: 'equation';
+        }
+        namespace Condition {
+            interface Equation {
+                /**
+                 * Left side of the equation
+                 */
+                left: string;
+                operator: '==' | '!=' | '>' | '>=' | '<' | '<=' | 'contains' | 'not_contains' | 'exists' | 'not_exist';
+                /**
+                 * Right side of the equation. The right side of the equation not required when
+                 * "exists" or "not_exist" are selected.
+                 */
+                right?: string;
+            }
+        }
+    }
+    interface SendSMSTool {
+        /**
+         * Name of the tool. Must be unique within all tools available to LLM at any given
+         * time (general tools + state tools + state edges).
+         */
+        name: string;
+        sms_content: SendSMSTool.SMSContentPredefined | SendSMSTool.SMSContentInferred | SendSMSTool.SMSContentTemplate;
+        type: 'send_sms';
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        condition?: SendSMSTool.Condition;
+        /**
+         * Names of tools that must run before this one.
+         */
+        depends_on?: Array<string>;
+        /**
+         * Describes what the tool does, sometimes can also include information about when
+         * to call the tool.
+         */
+        description?: string;
+        /**
+         * Describes what to say before sending the SMS. Only applicable when
+         * speak_during_execution is true.
+         */
+        execution_message_description?: string;
+        /**
+         * Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * If true, the agent will speak a short line before sending the SMS. If omitted,
+         * defaults to true (same as end_call / transfer_call tools).
+         */
+        speak_during_execution?: boolean;
+    }
+    namespace SendSMSTool {
+        interface SMSContentPredefined {
+            /**
+             * The static message to be sent in the SMS. Can contain dynamic variables.
+             */
+            text?: string;
+            type?: 'predefined';
+        }
+        interface SMSContentInferred {
+            /**
+             * The prompt to be used to help infer the SMS content. The model will take the
+             * global prompt, the call transcript, and this prompt together to deduce the right
+             * message to send. Can contain dynamic variables.
+             */
+            prompt?: string;
+            type?: 'inferred';
+        }
+        interface SMSContentTemplate {
+            /**
+             * The template to use for the SMS content. "info_collection" sends a predefined
+             * message requesting information from the user.
+             */
+            template: 'info_collection';
+            type: 'template';
+        }
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        interface Condition {
+            equations: Array<Condition.Equation>;
+            operator: '||' | '&&';
+            type: 'equation';
+        }
+        namespace Condition {
+            interface Equation {
+                /**
+                 * Left side of the equation
+                 */
+                left: string;
+                operator: '==' | '!=' | '>' | '>=' | '<' | '<=' | 'contains' | 'not_contains' | 'exists' | 'not_exist';
+                /**
+                 * Right side of the equation. The right side of the equation not required when
+                 * "exists" or "not_exist" are selected.
+                 */
+                right?: string;
+            }
+        }
+    }
+    interface AppTool {
+        /**
+         * The connection (App) this tool runs against. Must be a connection in the
+         * organization whose provider matches this tool's provider.
+         */
+        app_id: string;
+        /**
+         * Name of the catalog template within the provider, as listed by
+         * list-app-templates.
+         */
+        app_tool_template_name: string;
+        /**
+         * Name of the tool. Must be unique within the phase's tools; referenced by
+         * depends_on.
+         */
+        name: string;
+        /**
+         * Provider of the connection. Must match the connection's provider; supported
+         * providers are listed by list-app-templates.
+         */
+        provider: string;
+        type: 'integration_app';
+        /**
+         * Names of tools that must run before this one.
+         */
+        depends_on?: Array<string>;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Overrides the catalog template's LLM-facing description.
+         */
+        description?: string;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. If true, play a typing sound on the agent audio track while this
+         * tool is executing. Useful when the tool takes a noticeable amount of time to
+         * prevent silence on the call.
+         */
+        enable_typing_sound?: boolean;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. The message for the agent to speak when executing the tool. Only
+         * applicable when speak_during_execution is true.
+         */
+        execution_message_description?: string;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * What the agent and the transcript see of the tool's response. Omit to send the
+         * full response. Does not affect response_variables, which are always extracted
+         * from the raw response.
+         */
+        output_selection?: AppTool.UnionMember0 | AppTool.UnionMember1;
+        /**
+         * The resolved input parameters, in order. Properties may pin a value with const
+         * (including {{variable}} references) or provide a description for LLM inference.
+         * Each property may also record selected*input_mode, the editor mode the user
+         * selected ("const_enum", "const_boolean", "const_value", "description_custom", or
+         * "description_preset"); it is stored and returned as-is, used only by the tool
+         * config UI. Omit the key when no mode is recorded; when set, const*_ modes
+         * require a non-empty const, and description\__ modes must omit const entirely.
+         * Each parameter's required list must match the schema returned by the
+         * corresponding step of the get-app-tool-schema loop.
+         */
+        parameters?: Array<AppTool.Parameter>;
+        /**
+         * Mapping of a dynamic-variable name to the response field (dot-path) it is
+         * populated from. Missing paths are ignored.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Determines whether the agent would call LLM another time and speak
+         * when the result of the tool is obtained.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. If true, will speak during execution.
+         */
+        speak_during_execution?: boolean;
+    }
+    namespace AppTool {
+        interface UnionMember0 {
+            mode: 'all';
+            /**
+             * Not used at runtime; stored and returned as-is for the UI.
+             */
+            fields?: Array<string>;
+        }
+        interface UnionMember1 {
+            /**
+             * The only response fields the agent and the transcript see, as dot-paths into the
+             * response schema returned by get-app-tool-schema. Everything else is dropped.
+             * Selecting a parent keeps its whole subtree. A plain segment traverses arrays
+             * element-wise (deals.properties.amount keeps that field on every deal), while
+             * key[n] selects one element (deals[0].id keeps only the first deal's id); paths
+             * that match nothing contribute nothing.
+             */
+            fields: Array<string>;
+            mode: 'subset';
+        }
+        /**
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
+         */
+        interface Parameter {
+            /**
+             * The value of properties is an object, where each key is the name of a property
+             * and each value is a schema used to validate that property.
+             */
+            properties: unknown;
+            /**
+             * Type must be "object" for a JSON Schema object.
+             */
+            type: 'object';
+            /**
+             * List of names of required property when generating this parameter. LLM will do
+             * its best to generate the required properties in its function arguments. Property
+             * must exist in properties.
+             */
+            required?: Array<string>;
+        }
+    }
+    interface CustomTool {
+        /**
+         * Name of the tool. Must be unique within all tools available to LLM at any given
+         * time (general tools + state tools + state edges). Must be consisted of a-z, A-Z,
+         * 0-9, or contain underscores and dashes, with a maximum length of 64 (no space
+         * allowed).
+         */
+        name: string;
+        type: 'custom';
+        /**
+         * Describes what the tool does, sometimes can also include information about when
+         * to call the tool.
+         */
+        url: string;
+        /**
+         * If set to true, the parameters will be passed as root level JSON object instead
+         * of nested under "args".
+         */
+        args_at_root?: boolean;
+        /**
+         * Names of tools that must run before this one.
+         */
+        depends_on?: Array<string>;
+        /**
+         * Describes what this tool does and when to call this tool.
+         */
+        description?: string;
+        /**
+         * If true, play a typing sound on the agent audio track while this tool is
+         * executing. Useful when the tool takes a noticeable amount of time to prevent
+         * silence on the call.
+         */
+        enable_typing_sound?: boolean;
+        /**
+         * The description for the sentence agent say during execution. Only applicable
+         * when speak_during_execution is true. Can write what to say or even provide
+         * examples. The default is "The message you will say to callee when calling this
+         * tool. Make sure it fits into the conversation smoothly.".
+         */
+        execution_message_description?: string;
+        /**
+         * Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * Headers to add to the request.
+         */
+        headers?: {
+            [key: string]: string;
+        };
+        /**
+         * Maximum number of times to retry the request after a failed attempt, from 0 (no
+         * retry) to 5. Retries happen on any failure, with exponential backoff between
+         * attempts; the backoff delay is not configurable. `timeout_ms` applies per
+         * attempt rather than as a budget across all attempts, so an attempt that times
+         * out is still retried and the worst-case total duration is `timeout_ms`
+         * multiplied by (`max_retry` + 1) as well as any latency incurred by the
+         * exponential backoff + jitter between each retry. Only the final attempt's result
+         * is reported to the agent. Because retries repeat the request, only set this
+         * above 0 if your endpoint is idempotent — a retried request may be processed more
+         * than once. Defaults to 0 (no retry).
+         */
+        max_retry?: number;
+        /**
+         * Method to use for the request, default to POST.
+         */
+        method?: 'GET' | 'POST' | 'PUT' | 'PATCH' | 'DELETE';
+        /**
+         * How the tool's `parameters` are authored and shown in the dashboard editor —
+         * "form" for the visual parameter builder, "json" for a raw JSON Schema. Both
+         * produce the same `parameters` schema; this does not change how the request body
+         * is encoded (see `args_at_root`).
+         */
+        parameter_type?: 'json' | 'form';
+        /**
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
+         */
+        parameters?: CustomTool.Parameters;
+        /**
+         * Query parameters to append to the request URL.
+         */
+        query_params?: {
+            [key: string]: string;
+        };
+        /**
+         * A mapping of variable names to JSON paths in the response body. These values
+         * will be extracted from the response and made available as dynamic variables for
+         * use.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Determines whether the agent would call LLM another time and speak when the
+         * result of function is obtained. Usually this needs to get turned on so user can
+         * get update for the function call.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Determines whether the agent would say sentence like "One moment, let me check
+         * that." when executing the function. Recommend to turn on if your function call
+         * takes over 1s (including network) to complete, so that your agent remains
+         * responsive.
+         */
+        speak_during_execution?: boolean;
+        /**
+         * The maximum time in milliseconds the tool can run before it's considered
+         * timeout. If the tool times out, the agent would have that info. The minimum
+         * value allowed is 1000 ms (1 s), and maximum value allowed is 600,000 ms (10
+         * min). By default, this is set to 120,000 ms (2 min).
+         */
+        timeout_ms?: number;
+    }
+    namespace CustomTool {
+        /**
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
+         */
+        interface Parameters {
+            /**
+             * The value of properties is an object, where each key is the name of a property
+             * and each value is a schema used to validate that property.
+             */
+            properties: unknown;
+            /**
+             * Type must be "object" for a JSON Schema object.
+             */
+            type: 'object';
+            /**
+             * List of names of required property when generating this parameter. LLM will do
+             * its best to generate the required properties in its function arguments. Property
+             * must exist in properties.
+             */
+            required?: Array<string>;
+        }
+    }
+    interface CodeTool {
+        /**
+         * JavaScript code to execute in the sandbox.
+         */
+        code: string;
+        /**
+         * Name of the tool. Must be unique within all tools available to LLM at any given
+         * time (general tools + state tools + state edges). Must be consisted of a-z, A-Z,
+         * 0-9, or contain underscores and dashes, with a maximum length of 64 (no space
+         * allowed).
+         */
+        name: string;
+        type: 'code';
+        /**
+         * Names of tools that must run before this one.
+         */
+        depends_on?: Array<string>;
+        /**
+         * Describes what this tool does and when to call this tool.
+         */
+        description?: string;
+        /**
+         * If true, play a typing sound on the agent audio track while this tool is
+         * executing.
+         */
+        enable_typing_sound?: boolean;
+        /**
+         * The description for the sentence agent say during execution. Only applicable
+         * when speak_during_execution is true.
+         */
+        execution_message_description?: string;
+        /**
+         * Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * A mapping of variable names to JSON paths in the code execution result. These
+         * mapped values will be extracted and added as dynamic variables.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Determines whether the agent would call LLM another time and speak when the
+         * result of function is obtained.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Determines whether the agent would say sentence like "One moment, let me check
+         * that." when executing the tool.
+         */
+        speak_during_execution?: boolean;
+        /**
+         * The maximum time in milliseconds the code can run before it's considered
+         * timeout. Defaults to 30,000 ms (30 s).
+         */
+        timeout_ms?: number;
+    }
+}
+export interface ChatAgentListResponse {
+    /**
+     * Whether more results are available.
+     */
+    has_more: boolean;
+    items: Array<ChatAgentListResponse.Item>;
+    /**
+     * Pagination key for the next page.
+     */
+    pagination_key?: string;
+}
+export declare namespace ChatAgentListResponse {
+    interface Item {
+        /**
+         * Unique id of agent.
+         */
+        agent_id: string;
+        /**
+         * The name of the agent. Only used for your own reference.
+         */
+        agent_name: string;
+        channel: 'voice' | 'chat';
+        /**
+         * Authoritative root tags for this agent, keyed by tag name.
+         */
+        tags: {
+            [key: string]: Item.Tags;
+        };
+        /**
+         * User modification timestamp (milliseconds since epoch). Either the time of last
+         * update or creation if no updates available.
+         */
+        user_modified_timestamp: number;
+    }
+    namespace Item {
+        interface Tags {
+            dynamic_variables?: {
+                [key: string]: string;
+            };
+            version?: number;
+        }
+    }
+}
+export type ChatAgentCreateVersionResponse = AgentAPI.AgentResponse | ChatAgentResponse;
+export interface ChatAgentCreateParams {
+    /**
+     * The Response Engine to attach to the agent. It is used to generate responses for
+     * the agent. You need to create a Response Engine first before attaching it to an
+     * agent.
+     */
+    response_engine: ChatAgentCreateParams.ResponseEngineRetellLm | ChatAgentCreateParams.ResponseEngineCustomLm | ChatAgentCreateParams.ResponseEngineConversationFlow;
+    /**
+     * The name of the chat agent. Only used for your own reference.
+     */
+    agent_name?: string | null;
+    /**
+     * Message to display when the chat is automatically closed.
+     */
+    auto_close_message?: string | null;
+    /**
+     * Contact memory settings for phone calls and SMS chats. Creating an agent
+     * defaults enable_update to false and enable_read to true. Updates only change the
+     * supplied flags; omitted flags stay unchanged and an empty object has no effect.
+     * Set a flag to false to disable it. The configuration cannot be cleared. Existing
+     * agents without this configuration have both disabled.
+     */
+    contact_memory_config?: ChatAgentCreateParams.ContactMemoryConfig;
+    /**
+     * Number of days to retain call/chat data before automatic deletion. Must be
+     * between 1 and 730 days. If not set, data is retained forever (no automatic
+     * deletion).
+     */
+    data_storage_retention_days?: number | null;
+    /**
+     * Controls what data is stored for this agent. "everything" stores all data
+     * including transcripts and recordings. "everything_except_pii" stores data but
+     * excludes PII when possible based on PII configuration. "basic_attributes_only"
+     * stores only basic metadata. If not set, defaults to "everything".
+     */
+    data_storage_setting?: 'everything' | 'everything_except_pii' | 'basic_attributes_only' | null;
+    /**
+     * If users stay silent for a period after agent speech, end the chat. The minimum
+     * value allowed is 120,000 ms (2 minutes). The maximum value allowed is
+     * 259,200,000 ms (72 hours). By default, this is set to 3,600,000 (1 hour).
+     */
+    end_chat_after_silence_ms?: number | null;
+    /**
+     * Configuration for guardrail checks to detect and prevent prohibited topics in
+     * agent output and user input.
+     */
+    guardrail_config?: ChatAgentCreateParams.GuardrailConfig;
+    /**
+     * Toggle behavior presets on/off to influence agent response style and behaviors.
+     * Voice-only presets are not available for chat agents.
+     */
+    handbook_config?: ChatAgentCreateParams.HandbookConfig;
+    /**
+     * Specifies what language(s) the agent will operate in. Accepts either a single
+     * locale (e.g. `en-US`) or an array of locales for multilingual agents (e.g.
+     * `["en-US","es-ES"]`). The scalar value `multi` is deprecated but still accepted
+     * as a scalar, and is stored and returned as the ten locales it used to mean. It
+     * must not appear inside the array form. Send an explicit locale array instead. If
+     * unset, defaults to `en-US`.
+     */
+    language?: 'en-US' | 'en-IN' | 'en-GB' | 'en-AU' | 'en-NZ' | 'de-DE' | 'es-ES' | 'es-419' | 'hi-IN' | 'fr-FR' | 'fr-CA' | 'ja-JP' | 'pt-PT' | 'pt-BR' | 'zh-CN' | 'ru-RU' | 'it-IT' | 'ko-KR' | 'nl-NL' | 'nl-BE' | 'pl-PL' | 'tr-TR' | 'vi-VN' | 'ro-RO' | 'bg-BG' | 'ca-ES' | 'th-TH' | 'da-DK' | 'fi-FI' | 'el-GR' | 'hu-HU' | 'id-ID' | 'no-NO' | 'sk-SK' | 'sv-SE' | 'lt-LT' | 'lv-LV' | 'cs-CZ' | 'ms-MY' | 'af-ZA' | 'ar-SA' | 'az-AZ' | 'bs-BA' | 'cy-GB' | 'fa-IR' | 'fil-PH' | 'gl-ES' | 'he-IL' | 'hr-HR' | 'hy-AM' | 'is-IS' | 'kk-KZ' | 'kn-IN' | 'mk-MK' | 'mr-IN' | 'ne-NP' | 'sl-SI' | 'sr-RS' | 'sw-KE' | 'ta-IN' | 'ur-IN' | 'yue-CN' | 'uk-UA' | 'multi' | Array<'en-US' | 'en-IN' | 'en-GB' | 'en-AU' | 'en-NZ' | 'de-DE' | 'es-ES' | 'es-419' | 'hi-IN' | 'fr-FR' | 'fr-CA' | 'ja-JP' | 'pt-PT' | 'pt-BR' | 'zh-CN' | 'ru-RU' | 'it-IT' | 'ko-KR' | 'nl-NL' | 'nl-BE' | 'pl-PL' | 'tr-TR' | 'vi-VN' | 'ro-RO' | 'bg-BG' | 'ca-ES' | 'th-TH' | 'da-DK' | 'fi-FI' | 'el-GR' | 'hu-HU' | 'id-ID' | 'no-NO' | 'sk-SK' | 'sv-SE' | 'lt-LT' | 'lv-LV' | 'cs-CZ' | 'ms-MY' | 'af-ZA' | 'ar-SA' | 'az-AZ' | 'bs-BA' | 'cy-GB' | 'fa-IR' | 'fil-PH' | 'gl-ES' | 'he-IL' | 'hr-HR' | 'hy-AM' | 'is-IS' | 'kk-KZ' | 'kn-IN' | 'mk-MK' | 'mr-IN' | 'ne-NP' | 'sl-SI' | 'sr-RS' | 'sw-KE' | 'ta-IN' | 'ur-IN' | 'yue-CN' | 'uk-UA'>;
+    /**
+     * Whether this agent opts in to signed url for public log. If not set, default
+     * value of false will apply.
+     */
+    opt_in_signed_url?: boolean;
+    /**
+     * Configuration for PII scrubbing from transcripts and recordings.
+     */
+    pii_config?: ChatAgentCreateParams.PiiConfig;
+    /**
+     * Post chat analysis data to extract from the chat. This data will augment the
+     * pre-defined variables extracted in the chat analysis. This will be available
+     * after the chat ends.
+     */
+    post_chat_analysis_data?: Array<ChatAgentCreateParams.StringAnalysisData | ChatAgentCreateParams.EnumAnalysisData | ChatAgentCreateParams.BooleanAnalysisData | ChatAgentCreateParams.NumberAnalysisData | ChatAgentCreateParams.ChatPresetAnalysisData> | null;
+    /**
+     * The model to use for post chat analysis. Default to gpt-5.6-terra.
+     */
+    post_chat_analysis_model?: 'gpt-4.1' | 'gpt-4.1-mini' | 'gpt-4.1-nano' | 'gpt-5' | 'gpt-5-mini' | 'gpt-5-nano' | 'gpt-5.1' | 'gpt-5.2' | 'gpt-5.4' | 'gpt-5.4-mini' | 'gpt-5.4-nano' | 'gpt-5.5' | 'gpt-5.6-terra' | 'gpt-5.6-luna' | 'gpt-6-astra' | 'gpt-6-sol' | 'gpt-6.1-sol' | 'gpt-6-luna' | 'claude-4.5-sonnet' | 'claude-4.6-sonnet' | 'claude-5-opus' | 'claude-5.5-opus' | 'claude-5-sonnet' | 'claude-5.5-sonnet' | 'claude-4.5-haiku' | 'gemini-3.0-flash' | 'gemini-3.1-flash-lite' | 'gemini-3.5-flash' | 'gemini-3.5-flash-lite' | 'gemini-3.6-flash' | 'gemini-3.7-flash' | 'gemini-3.8-flash' | null;
+    /**
+     * Integration (Agent Functions) tools run as a dependency graph at chat end, after
+     * post-chat analysis. Each tool can be gated by a condition. Set to null to clear.
+     */
+    post_session_tools?: Array<ChatAgentCreateParams.AppTool | ChatAgentCreateParams.CustomTool | ChatAgentCreateParams.CodeTool | ChatAgentCreateParams.SendSMSTool> | null;
+    /**
+     * Integration (Agent Functions) tools run as a dependency graph before the chat's
+     * first message. Outputs are injected as dynamic variables. Set to null to clear.
+     */
+    pre_session_tools?: Array<ChatAgentCreateParams.AppTool | ChatAgentCreateParams.CustomTool | ChatAgentCreateParams.CodeTool> | null;
+    /**
+     * The expiration time for the signed url in milliseconds. Only applicable when
+     * opt_in_signed_url is true. If not set, default value of 86400000 (24 hours) will
+     * apply.
+     */
+    signed_url_expiration_ms?: number | null;
+    /**
+     * IANA timezone for the agent (e.g. America/New_York). Defaults to
+     * America/Los_Angeles if not set.
+     */
+    timezone?: string | null;
+    /**
+     * Optional title of the chat agent version. Used for your own reference.
+     */
+    version_title?: string | null;
+    /**
+     * Which webhook events this agent should receive. If not set, defaults to
+     * chat_started, chat_ended, chat_analyzed.
+     */
+    webhook_events?: Array<'chat_started' | 'chat_ended' | 'chat_analyzed' | 'transcript_updated'> | null;
+    /**
+     * The timeout for the webhook in milliseconds. If not set, default value of 10000
+     * will apply.
+     */
+    webhook_timeout_ms?: number;
+    /**
+     * The webhook for agent to listen to chat events. See what events it would get at
+     * [webhook doc](/features/webhook). If set, will binds webhook events for this
+     * agent to the specified url, and will ignore the account level webhook for this
+     * agent. Set to `null` to remove webhook url from this agent.
+     */
+    webhook_url?: string | null;
+}
+export declare namespace ChatAgentCreateParams {
+    interface ResponseEngineRetellLm {
+        /**
+         * id of the Retell LLM Response Engine.
+         */
+        llm_id: string;
+        /**
+         * type of the Response Engine.
+         */
+        type: 'retell-llm';
+        /**
+         * Version of the Retell LLM Response Engine.
+         */
+        version?: number | null;
+    }
+    interface ResponseEngineCustomLm {
+        /**
+         * LLM websocket url of the custom LLM.
+         */
+        llm_websocket_url: string;
+        /**
+         * type of the Response Engine.
+         */
+        type: 'custom-llm';
+    }
+    interface ResponseEngineConversationFlow {
+        /**
+         * ID of the Conversation Flow Response Engine.
+         */
+        conversation_flow_id: string;
+        /**
+         * type of the Response Engine.
+         */
+        type: 'conversation-flow';
+        /**
+         * Version of the Conversation Flow Response Engine.
+         */
+        version?: number | null;
+    }
+    /**
+     * Contact memory settings for phone calls and SMS chats. Creating an agent
+     * defaults enable_update to false and enable_read to true. Updates only change the
+     * supplied flags; omitted flags stay unchanged and an empty object has no effect.
+     * Set a flag to false to disable it. The configuration cannot be cleared. Existing
+     * agents without this configuration have both disabled.
+     */
+    interface ContactMemoryConfig {
+        /**
+         * Automatically add saved contact memory to the agent prompt. Skippable nodes can
+         * use answers from the current conversation even when this setting is disabled.
+         * Contact dynamic variables, including contact_memory, remain available regardless
+         * of this setting.
+         */
+        enable_read?: boolean;
+        /**
+         * Rewrite the contact memory after each conversation. Requires storing
+         * conversation data. Chat agents must also have end_chat_after_silence_ms set.
+         */
+        enable_update?: boolean;
+    }
+    /**
+     * Configuration for guardrail checks to detect and prevent prohibited topics in
+     * agent output and user input.
+     */
+    interface GuardrailConfig {
+        /**
+         * Selected prohibited user topic categories to check. When user messages contain
+         * these topics, the agent will respond with a placeholder message instead of
+         * processing the request.
+         */
+        input_topics?: Array<'platform_integrity_jailbreaking'> | null;
+        /**
+         * Selected prohibited agent topic categories to check. When agent messages contain
+         * these topics, they will be replaced with a placeholder message.
+         */
+        output_topics?: Array<'harassment' | 'self_harm' | 'sexual_exploitation' | 'violence' | 'defense_and_national_security' | 'illicit_and_harmful_activity' | 'gambling' | 'regulated_professional_advice' | 'child_safety_and_exploitation'> | null;
+    }
+    /**
+     * Toggle behavior presets on/off to influence agent response style and behaviors.
+     * Voice-only presets are not available for chat agents.
+     */
+    interface HandbookConfig {
+        /**
+         * When asked, acknowledge being a virtual assistant.
+         */
+        ai_disclosure?: boolean;
+        /**
+         * Professional call center rep baseline.
+         */
+        default_personality?: boolean;
+        /**
+         * Warm acknowledgment of caller concerns.
+         */
+        high_empathy?: boolean;
+        /**
+         * Stay within prompt/context scope, don't invent details.
+         */
+        scope_boundaries?: boolean;
+    }
+    /**
+     * Configuration for PII scrubbing from transcripts and recordings.
+     */
+    interface PiiConfig {
+        /**
+         * List of PII categories to scrub from transcripts and recordings. PII redaction
+         * is only active when this list is non-empty; an empty array means no PII
+         * scrubbing is performed.
+         */
+        categories: Array<'person_name' | 'address' | 'email' | 'phone_number' | 'ssn' | 'passport' | 'driver_license' | 'credit_card' | 'bank_account' | 'password' | 'pin' | 'medical_id' | 'date_of_birth' | 'customer_account_number'>;
+        /**
+         * The processing mode for PII scrubbing. Currently only post-call is supported.
+         */
+        mode: 'post_call';
+    }
+    interface StringAnalysisData {
+        /**
+         * Description of the variable.
+         */
+        description: string;
+        /**
+         * Name of the variable.
+         */
+        name: string;
+        /**
+         * Type of the variable to extract.
+         */
+        type: 'string';
+        /**
+         * Optional instruction to help decide whether this field needs to be populated in
+         * the analysis. If not set, the field is always included. If required is true,
+         * this is ignored.
+         */
+        conditional_prompt?: string;
+        /**
+         * Examples of the variable value to teach model the style and syntax.
+         */
+        examples?: Array<string>;
+        /**
+         * Whether this data is required. If true and the data is not extracted, the call
+         * will be marked as unsuccessful.
+         */
+        required?: boolean;
+    }
+    interface EnumAnalysisData {
+        /**
+         * The possible values of the variable, must be non empty array.
+         */
+        choices: Array<string>;
+        /**
+         * Description of the variable.
+         */
+        description: string;
+        /**
+         * Name of the variable.
+         */
+        name: string;
+        /**
+         * Type of the variable to extract.
+         */
+        type: 'enum';
+        /**
+         * Optional instruction to help decide whether this field needs to be populated in
+         * the analysis. If not set, the field is always included. If required is true,
+         * this is ignored.
+         */
+        conditional_prompt?: string;
+        /**
+         * Whether this data is required. If true and the data is not extracted, the call
+         * will be marked as unsuccessful.
+         */
+        required?: boolean;
+    }
+    interface BooleanAnalysisData {
+        /**
+         * Description of the variable.
+         */
+        description: string;
+        /**
+         * Name of the variable.
+         */
+        name: string;
+        /**
+         * Type of the variable to extract.
+         */
+        type: 'boolean';
+        /**
+         * Optional instruction to help decide whether this field needs to be populated in
+         * the analysis. If not set, the field is always included. If required is true,
+         * this is ignored.
+         */
+        conditional_prompt?: string;
+        /**
+         * Whether this data is required. If true and the data is not extracted, the call
+         * will be marked as unsuccessful.
+         */
+        required?: boolean;
+    }
+    interface NumberAnalysisData {
+        /**
+         * Description of the variable.
+         */
+        description: string;
+        /**
+         * Name of the variable.
+         */
+        name: string;
+        /**
+         * Type of the variable to extract.
+         */
+        type: 'number';
+        /**
+         * Optional instruction to help decide whether this field needs to be populated in
+         * the analysis. If not set, the field is always included. If required is true,
+         * this is ignored.
+         */
+        conditional_prompt?: string;
+        /**
+         * Whether this data is required. If true and the data is not extracted, the call
+         * will be marked as unsuccessful.
+         */
+        required?: boolean;
+    }
+    /**
+     * System preset for post-chat analysis (chat agents). Use in
+     * post_chat_analysis_data to override prompts or mark fields optional.
+     */
+    interface ChatPresetAnalysisData {
+        /**
+         * Preset identifier for chat agent analysis.
+         */
+        name: 'chat_summary' | 'chat_successful' | 'user_sentiment';
+        /**
+         * Identifies this item as a system preset.
+         */
+        type: 'system-presets';
+        /**
+         * Optional instruction to help decide whether this field needs to be populated. If
+         * not set, the field is always included.
+         */
+        conditional_prompt?: string;
+        /**
+         * Prompt or description for this preset.
+         */
+        description?: string;
+        /**
+         * If false, this field is optional in the analysis. If true or unset, the field is
+         * required.
+         */
+        required?: boolean;
+    }
+    interface AppTool {
+        /**
+         * The connection (App) this tool runs against. Must be a connection in the
+         * organization whose provider matches this tool's provider.
+         */
+        app_id: string;
+        /**
+         * Name of the catalog template within the provider, as listed by
+         * list-app-templates.
+         */
+        app_tool_template_name: string;
+        /**
+         * Name of the tool. Must be unique within the phase's tools; referenced by
+         * depends_on.
+         */
+        name: string;
+        /**
+         * Provider of the connection. Must match the connection's provider; supported
+         * providers are listed by list-app-templates.
+         */
+        provider: string;
+        type: 'integration_app';
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        condition?: AppTool.Condition;
+        /**
+         * Names of tools that must run before this one.
+         */
+        depends_on?: Array<string>;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Overrides the catalog template's LLM-facing description.
+         */
+        description?: string;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. If true, play a typing sound on the agent audio track while this
+         * tool is executing. Useful when the tool takes a noticeable amount of time to
+         * prevent silence on the call.
+         */
+        enable_typing_sound?: boolean;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. The message for the agent to speak when executing the tool. Only
+         * applicable when speak_during_execution is true.
+         */
+        execution_message_description?: string;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * What the agent and the transcript see of the tool's response. Omit to send the
+         * full response. Does not affect response_variables, which are always extracted
+         * from the raw response.
+         */
+        output_selection?: AppTool.UnionMember0 | AppTool.UnionMember1;
+        /**
+         * The resolved input parameters, in order. Properties may pin a value with const
+         * (including {{variable}} references) or provide a description for LLM inference.
+         * Each property may also record selected*input_mode, the editor mode the user
+         * selected ("const_enum", "const_boolean", "const_value", "description_custom", or
+         * "description_preset"); it is stored and returned as-is, used only by the tool
+         * config UI. Omit the key when no mode is recorded; when set, const*_ modes
+         * require a non-empty const, and description\__ modes must omit const entirely.
+         * Each parameter's required list must match the schema returned by the
+         * corresponding step of the get-app-tool-schema loop.
+         */
+        parameters?: Array<AppTool.Parameter>;
+        /**
+         * Mapping of a dynamic-variable name to the response field (dot-path) it is
+         * populated from. Missing paths are ignored.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Determines whether the agent would call LLM another time and speak
+         * when the result of the tool is obtained.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. If true, will speak during execution.
+         */
+        speak_during_execution?: boolean;
+    }
+    namespace AppTool {
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        interface Condition {
+            equations: Array<Condition.Equation>;
+            operator: '||' | '&&';
+            type: 'equation';
+        }
+        namespace Condition {
+            interface Equation {
+                /**
+                 * Left side of the equation
+                 */
+                left: string;
+                operator: '==' | '!=' | '>' | '>=' | '<' | '<=' | 'contains' | 'not_contains' | 'exists' | 'not_exist';
+                /**
+                 * Right side of the equation. The right side of the equation not required when
+                 * "exists" or "not_exist" are selected.
+                 */
+                right?: string;
+            }
+        }
+        interface UnionMember0 {
+            mode: 'all';
+            /**
+             * Not used at runtime; stored and returned as-is for the UI.
+             */
+            fields?: Array<string>;
+        }
+        interface UnionMember1 {
+            /**
+             * The only response fields the agent and the transcript see, as dot-paths into the
+             * response schema returned by get-app-tool-schema. Everything else is dropped.
+             * Selecting a parent keeps its whole subtree. A plain segment traverses arrays
+             * element-wise (deals.properties.amount keeps that field on every deal), while
+             * key[n] selects one element (deals[0].id keeps only the first deal's id); paths
+             * that match nothing contribute nothing.
+             */
+            fields: Array<string>;
+            mode: 'subset';
+        }
+        /**
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
+         */
+        interface Parameter {
+            /**
+             * The value of properties is an object, where each key is the name of a property
+             * and each value is a schema used to validate that property.
+             */
+            properties: unknown;
+            /**
+             * Type must be "object" for a JSON Schema object.
+             */
+            type: 'object';
+            /**
+             * List of names of required property when generating this parameter. LLM will do
+             * its best to generate the required properties in its function arguments. Property
+             * must exist in properties.
+             */
+            required?: Array<string>;
+        }
+    }
+    interface CustomTool {
+        /**
+         * Name of the tool. Must be unique within all tools available to LLM at any given
+         * time (general tools + state tools + state edges). Must be consisted of a-z, A-Z,
+         * 0-9, or contain underscores and dashes, with a maximum length of 64 (no space
+         * allowed).
+         */
+        name: string;
+        type: 'custom';
+        /**
+         * Describes what the tool does, sometimes can also include information about when
+         * to call the tool.
+         */
+        url: string;
+        /**
+         * If set to true, the parameters will be passed as root level JSON object instead
+         * of nested under "args".
+         */
+        args_at_root?: boolean;
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        condition?: CustomTool.Condition;
+        /**
+         * Names of tools that must run before this one.
+         */
+        depends_on?: Array<string>;
+        /**
+         * Describes what this tool does and when to call this tool.
+         */
+        description?: string;
+        /**
+         * If true, play a typing sound on the agent audio track while this tool is
+         * executing. Useful when the tool takes a noticeable amount of time to prevent
+         * silence on the call.
+         */
+        enable_typing_sound?: boolean;
+        /**
+         * The description for the sentence agent say during execution. Only applicable
+         * when speak_during_execution is true. Can write what to say or even provide
+         * examples. The default is "The message you will say to callee when calling this
+         * tool. Make sure it fits into the conversation smoothly.".
+         */
+        execution_message_description?: string;
+        /**
+         * Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * Headers to add to the request.
+         */
+        headers?: {
+            [key: string]: string;
+        };
+        /**
+         * Maximum number of times to retry the request after a failed attempt, from 0 (no
+         * retry) to 5. Retries happen on any failure, with exponential backoff between
+         * attempts; the backoff delay is not configurable. `timeout_ms` applies per
+         * attempt rather than as a budget across all attempts, so an attempt that times
+         * out is still retried and the worst-case total duration is `timeout_ms`
+         * multiplied by (`max_retry` + 1) as well as any latency incurred by the
+         * exponential backoff + jitter between each retry. Only the final attempt's result
+         * is reported to the agent. Because retries repeat the request, only set this
+         * above 0 if your endpoint is idempotent — a retried request may be processed more
+         * than once. Defaults to 0 (no retry).
+         */
+        max_retry?: number;
+        /**
+         * Method to use for the request, default to POST.
+         */
+        method?: 'GET' | 'POST' | 'PUT' | 'PATCH' | 'DELETE';
+        /**
+         * How the tool's `parameters` are authored and shown in the dashboard editor —
+         * "form" for the visual parameter builder, "json" for a raw JSON Schema. Both
+         * produce the same `parameters` schema; this does not change how the request body
+         * is encoded (see `args_at_root`).
+         */
+        parameter_type?: 'json' | 'form';
+        /**
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
+         */
+        parameters?: CustomTool.Parameters;
+        /**
+         * Query parameters to append to the request URL.
+         */
+        query_params?: {
+            [key: string]: string;
+        };
+        /**
+         * A mapping of variable names to JSON paths in the response body. These values
+         * will be extracted from the response and made available as dynamic variables for
+         * use.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Determines whether the agent would call LLM another time and speak when the
+         * result of function is obtained. Usually this needs to get turned on so user can
+         * get update for the function call.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Determines whether the agent would say sentence like "One moment, let me check
+         * that." when executing the function. Recommend to turn on if your function call
+         * takes over 1s (including network) to complete, so that your agent remains
+         * responsive.
+         */
+        speak_during_execution?: boolean;
+        /**
+         * The maximum time in milliseconds the tool can run before it's considered
+         * timeout. If the tool times out, the agent would have that info. The minimum
+         * value allowed is 1000 ms (1 s), and maximum value allowed is 600,000 ms (10
+         * min). By default, this is set to 120,000 ms (2 min).
+         */
+        timeout_ms?: number;
+    }
+    namespace CustomTool {
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        interface Condition {
+            equations: Array<Condition.Equation>;
+            operator: '||' | '&&';
+            type: 'equation';
+        }
+        namespace Condition {
+            interface Equation {
+                /**
+                 * Left side of the equation
+                 */
+                left: string;
+                operator: '==' | '!=' | '>' | '>=' | '<' | '<=' | 'contains' | 'not_contains' | 'exists' | 'not_exist';
+                /**
+                 * Right side of the equation. The right side of the equation not required when
+                 * "exists" or "not_exist" are selected.
+                 */
+                right?: string;
+            }
+        }
+        /**
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
+         */
+        interface Parameters {
+            /**
+             * The value of properties is an object, where each key is the name of a property
+             * and each value is a schema used to validate that property.
+             */
+            properties: unknown;
+            /**
+             * Type must be "object" for a JSON Schema object.
+             */
+            type: 'object';
+            /**
+             * List of names of required property when generating this parameter. LLM will do
+             * its best to generate the required properties in its function arguments. Property
+             * must exist in properties.
+             */
+            required?: Array<string>;
+        }
+    }
+    interface CodeTool {
+        /**
+         * JavaScript code to execute in the sandbox.
+         */
+        code: string;
+        /**
+         * Name of the tool. Must be unique within all tools available to LLM at any given
+         * time (general tools + state tools + state edges). Must be consisted of a-z, A-Z,
+         * 0-9, or contain underscores and dashes, with a maximum length of 64 (no space
+         * allowed).
+         */
+        name: string;
+        type: 'code';
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        condition?: CodeTool.Condition;
+        /**
+         * Names of tools that must run before this one.
+         */
+        depends_on?: Array<string>;
+        /**
+         * Describes what this tool does and when to call this tool.
+         */
+        description?: string;
+        /**
+         * If true, play a typing sound on the agent audio track while this tool is
+         * executing.
+         */
+        enable_typing_sound?: boolean;
+        /**
+         * The description for the sentence agent say during execution. Only applicable
+         * when speak_during_execution is true.
+         */
+        execution_message_description?: string;
+        /**
+         * Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * A mapping of variable names to JSON paths in the code execution result. These
+         * mapped values will be extracted and added as dynamic variables.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Determines whether the agent would call LLM another time and speak when the
+         * result of function is obtained.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Determines whether the agent would say sentence like "One moment, let me check
+         * that." when executing the tool.
+         */
+        speak_during_execution?: boolean;
+        /**
+         * The maximum time in milliseconds the code can run before it's considered
+         * timeout. Defaults to 30,000 ms (30 s).
+         */
+        timeout_ms?: number;
+    }
+    namespace CodeTool {
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        interface Condition {
+            equations: Array<Condition.Equation>;
+            operator: '||' | '&&';
+            type: 'equation';
+        }
+        namespace Condition {
+            interface Equation {
+                /**
+                 * Left side of the equation
+                 */
+                left: string;
+                operator: '==' | '!=' | '>' | '>=' | '<' | '<=' | 'contains' | 'not_contains' | 'exists' | 'not_exist';
+                /**
+                 * Right side of the equation. The right side of the equation not required when
+                 * "exists" or "not_exist" are selected.
+                 */
+                right?: string;
+            }
+        }
+    }
+    interface SendSMSTool {
+        /**
+         * Name of the tool. Must be unique within all tools available to LLM at any given
+         * time (general tools + state tools + state edges).
+         */
+        name: string;
+        sms_content: SendSMSTool.SMSContentPredefined | SendSMSTool.SMSContentInferred | SendSMSTool.SMSContentTemplate;
+        type: 'send_sms';
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        condition?: SendSMSTool.Condition;
+        /**
+         * Names of tools that must run before this one.
+         */
+        depends_on?: Array<string>;
+        /**
+         * Describes what the tool does, sometimes can also include information about when
+         * to call the tool.
+         */
+        description?: string;
+        /**
+         * Describes what to say before sending the SMS. Only applicable when
+         * speak_during_execution is true.
+         */
+        execution_message_description?: string;
+        /**
+         * Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * If true, the agent will speak a short line before sending the SMS. If omitted,
+         * defaults to true (same as end_call / transfer_call tools).
+         */
+        speak_during_execution?: boolean;
+    }
+    namespace SendSMSTool {
+        interface SMSContentPredefined {
+            /**
+             * The static message to be sent in the SMS. Can contain dynamic variables.
+             */
+            text?: string;
+            type?: 'predefined';
+        }
+        interface SMSContentInferred {
+            /**
+             * The prompt to be used to help infer the SMS content. The model will take the
+             * global prompt, the call transcript, and this prompt together to deduce the right
+             * message to send. Can contain dynamic variables.
+             */
+            prompt?: string;
+            type?: 'inferred';
+        }
+        interface SMSContentTemplate {
+            /**
+             * The template to use for the SMS content. "info_collection" sends a predefined
+             * message requesting information from the user.
+             */
+            template: 'info_collection';
+            type: 'template';
+        }
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        interface Condition {
+            equations: Array<Condition.Equation>;
+            operator: '||' | '&&';
+            type: 'equation';
+        }
+        namespace Condition {
+            interface Equation {
+                /**
+                 * Left side of the equation
+                 */
+                left: string;
+                operator: '==' | '!=' | '>' | '>=' | '<' | '<=' | 'contains' | 'not_contains' | 'exists' | 'not_exist';
+                /**
+                 * Right side of the equation. The right side of the equation not required when
+                 * "exists" or "not_exist" are selected.
+                 */
+                right?: string;
+            }
+        }
+    }
+    interface AppTool {
+        /**
+         * The connection (App) this tool runs against. Must be a connection in the
+         * organization whose provider matches this tool's provider.
+         */
+        app_id: string;
+        /**
+         * Name of the catalog template within the provider, as listed by
+         * list-app-templates.
+         */
+        app_tool_template_name: string;
+        /**
+         * Name of the tool. Must be unique within the phase's tools; referenced by
+         * depends_on.
+         */
+        name: string;
+        /**
+         * Provider of the connection. Must match the connection's provider; supported
+         * providers are listed by list-app-templates.
+         */
+        provider: string;
+        type: 'integration_app';
+        /**
+         * Names of tools that must run before this one.
+         */
+        depends_on?: Array<string>;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Overrides the catalog template's LLM-facing description.
+         */
+        description?: string;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. If true, play a typing sound on the agent audio track while this
+         * tool is executing. Useful when the tool takes a noticeable amount of time to
+         * prevent silence on the call.
+         */
+        enable_typing_sound?: boolean;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. The message for the agent to speak when executing the tool. Only
+         * applicable when speak_during_execution is true.
+         */
+        execution_message_description?: string;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * What the agent and the transcript see of the tool's response. Omit to send the
+         * full response. Does not affect response_variables, which are always extracted
+         * from the raw response.
+         */
+        output_selection?: AppTool.UnionMember0 | AppTool.UnionMember1;
+        /**
+         * The resolved input parameters, in order. Properties may pin a value with const
+         * (including {{variable}} references) or provide a description for LLM inference.
+         * Each property may also record selected*input_mode, the editor mode the user
+         * selected ("const_enum", "const_boolean", "const_value", "description_custom", or
+         * "description_preset"); it is stored and returned as-is, used only by the tool
+         * config UI. Omit the key when no mode is recorded; when set, const*_ modes
+         * require a non-empty const, and description\__ modes must omit const entirely.
+         * Each parameter's required list must match the schema returned by the
+         * corresponding step of the get-app-tool-schema loop.
+         */
+        parameters?: Array<AppTool.Parameter>;
+        /**
+         * Mapping of a dynamic-variable name to the response field (dot-path) it is
+         * populated from. Missing paths are ignored.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Determines whether the agent would call LLM another time and speak
+         * when the result of the tool is obtained.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. If true, will speak during execution.
+         */
+        speak_during_execution?: boolean;
+    }
+    namespace AppTool {
+        interface UnionMember0 {
+            mode: 'all';
+            /**
+             * Not used at runtime; stored and returned as-is for the UI.
+             */
+            fields?: Array<string>;
+        }
+        interface UnionMember1 {
+            /**
+             * The only response fields the agent and the transcript see, as dot-paths into the
+             * response schema returned by get-app-tool-schema. Everything else is dropped.
+             * Selecting a parent keeps its whole subtree. A plain segment traverses arrays
+             * element-wise (deals.properties.amount keeps that field on every deal), while
+             * key[n] selects one element (deals[0].id keeps only the first deal's id); paths
+             * that match nothing contribute nothing.
+             */
+            fields: Array<string>;
+            mode: 'subset';
+        }
+        /**
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
+         */
+        interface Parameter {
+            /**
+             * The value of properties is an object, where each key is the name of a property
+             * and each value is a schema used to validate that property.
+             */
+            properties: unknown;
+            /**
+             * Type must be "object" for a JSON Schema object.
+             */
+            type: 'object';
+            /**
+             * List of names of required property when generating this parameter. LLM will do
+             * its best to generate the required properties in its function arguments. Property
+             * must exist in properties.
+             */
+            required?: Array<string>;
+        }
+    }
+    interface CustomTool {
+        /**
+         * Name of the tool. Must be unique within all tools available to LLM at any given
+         * time (general tools + state tools + state edges). Must be consisted of a-z, A-Z,
+         * 0-9, or contain underscores and dashes, with a maximum length of 64 (no space
+         * allowed).
+         */
+        name: string;
+        type: 'custom';
+        /**
+         * Describes what the tool does, sometimes can also include information about when
+         * to call the tool.
+         */
+        url: string;
+        /**
+         * If set to true, the parameters will be passed as root level JSON object instead
+         * of nested under "args".
+         */
+        args_at_root?: boolean;
+        /**
+         * Names of tools that must run before this one.
+         */
+        depends_on?: Array<string>;
+        /**
+         * Describes what this tool does and when to call this tool.
+         */
+        description?: string;
+        /**
+         * If true, play a typing sound on the agent audio track while this tool is
+         * executing. Useful when the tool takes a noticeable amount of time to prevent
+         * silence on the call.
+         */
+        enable_typing_sound?: boolean;
+        /**
+         * The description for the sentence agent say during execution. Only applicable
+         * when speak_during_execution is true. Can write what to say or even provide
+         * examples. The default is "The message you will say to callee when calling this
+         * tool. Make sure it fits into the conversation smoothly.".
+         */
+        execution_message_description?: string;
+        /**
+         * Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * Headers to add to the request.
+         */
+        headers?: {
+            [key: string]: string;
+        };
+        /**
+         * Maximum number of times to retry the request after a failed attempt, from 0 (no
+         * retry) to 5. Retries happen on any failure, with exponential backoff between
+         * attempts; the backoff delay is not configurable. `timeout_ms` applies per
+         * attempt rather than as a budget across all attempts, so an attempt that times
+         * out is still retried and the worst-case total duration is `timeout_ms`
+         * multiplied by (`max_retry` + 1) as well as any latency incurred by the
+         * exponential backoff + jitter between each retry. Only the final attempt's result
+         * is reported to the agent. Because retries repeat the request, only set this
+         * above 0 if your endpoint is idempotent — a retried request may be processed more
+         * than once. Defaults to 0 (no retry).
+         */
+        max_retry?: number;
+        /**
+         * Method to use for the request, default to POST.
+         */
+        method?: 'GET' | 'POST' | 'PUT' | 'PATCH' | 'DELETE';
+        /**
+         * How the tool's `parameters` are authored and shown in the dashboard editor —
+         * "form" for the visual parameter builder, "json" for a raw JSON Schema. Both
+         * produce the same `parameters` schema; this does not change how the request body
+         * is encoded (see `args_at_root`).
+         */
+        parameter_type?: 'json' | 'form';
+        /**
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
+         */
+        parameters?: CustomTool.Parameters;
+        /**
+         * Query parameters to append to the request URL.
+         */
+        query_params?: {
+            [key: string]: string;
+        };
+        /**
+         * A mapping of variable names to JSON paths in the response body. These values
+         * will be extracted from the response and made available as dynamic variables for
+         * use.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Determines whether the agent would call LLM another time and speak when the
+         * result of function is obtained. Usually this needs to get turned on so user can
+         * get update for the function call.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Determines whether the agent would say sentence like "One moment, let me check
+         * that." when executing the function. Recommend to turn on if your function call
+         * takes over 1s (including network) to complete, so that your agent remains
+         * responsive.
+         */
+        speak_during_execution?: boolean;
+        /**
+         * The maximum time in milliseconds the tool can run before it's considered
+         * timeout. If the tool times out, the agent would have that info. The minimum
+         * value allowed is 1000 ms (1 s), and maximum value allowed is 600,000 ms (10
+         * min). By default, this is set to 120,000 ms (2 min).
+         */
+        timeout_ms?: number;
+    }
+    namespace CustomTool {
+        /**
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
+         */
+        interface Parameters {
+            /**
+             * The value of properties is an object, where each key is the name of a property
+             * and each value is a schema used to validate that property.
+             */
+            properties: unknown;
+            /**
+             * Type must be "object" for a JSON Schema object.
+             */
+            type: 'object';
+            /**
+             * List of names of required property when generating this parameter. LLM will do
+             * its best to generate the required properties in its function arguments. Property
+             * must exist in properties.
+             */
+            required?: Array<string>;
+        }
+    }
+    interface CodeTool {
+        /**
+         * JavaScript code to execute in the sandbox.
+         */
+        code: string;
+        /**
+         * Name of the tool. Must be unique within all tools available to LLM at any given
+         * time (general tools + state tools + state edges). Must be consisted of a-z, A-Z,
+         * 0-9, or contain underscores and dashes, with a maximum length of 64 (no space
+         * allowed).
+         */
+        name: string;
+        type: 'code';
+        /**
+         * Names of tools that must run before this one.
+         */
+        depends_on?: Array<string>;
+        /**
+         * Describes what this tool does and when to call this tool.
+         */
+        description?: string;
+        /**
+         * If true, play a typing sound on the agent audio track while this tool is
+         * executing.
+         */
+        enable_typing_sound?: boolean;
+        /**
+         * The description for the sentence agent say during execution. Only applicable
+         * when speak_during_execution is true.
+         */
+        execution_message_description?: string;
+        /**
+         * Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * A mapping of variable names to JSON paths in the code execution result. These
+         * mapped values will be extracted and added as dynamic variables.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Determines whether the agent would call LLM another time and speak when the
+         * result of function is obtained.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Determines whether the agent would say sentence like "One moment, let me check
+         * that." when executing the tool.
+         */
+        speak_during_execution?: boolean;
+        /**
+         * The maximum time in milliseconds the code can run before it's considered
+         * timeout. Defaults to 30,000 ms (30 s).
+         */
+        timeout_ms?: number;
+    }
+}
+export interface ChatAgentRetrieveParams {
+    /**
+     * Optional version of the API to use for this request. If not provided, will
+     * default to latest version.
+     */
+    version?: string | number;
+}
+export interface ChatAgentUpdateParams {
+    /**
+     * Query param: Optional version of the API to use for this request. Default to
+     * latest version.
+     */
+    version?: string | number;
+    /**
+     * Body param: The name of the chat agent. Only used for your own reference.
+     */
+    agent_name?: string | null;
+    /**
+     * Body param: Message to display when the chat is automatically closed.
+     */
+    auto_close_message?: string | null;
+    /**
+     * Body param: Contact memory settings for phone calls and SMS chats. Creating an
+     * agent defaults enable_update to false and enable_read to true. Updates only
+     * change the supplied flags; omitted flags stay unchanged and an empty object has
+     * no effect. Set a flag to false to disable it. The configuration cannot be
+     * cleared. Existing agents without this configuration have both disabled.
+     */
+    contact_memory_config?: ChatAgentUpdateParams.ContactMemoryConfig;
+    /**
+     * Body param: Number of days to retain call/chat data before automatic deletion.
+     * Must be between 1 and 730 days. If not set, data is retained forever (no
+     * automatic deletion).
+     */
+    data_storage_retention_days?: number | null;
+    /**
+     * Body param: Controls what data is stored for this agent. "everything" stores all
+     * data including transcripts and recordings. "everything_except_pii" stores data
+     * but excludes PII when possible based on PII configuration.
+     * "basic_attributes_only" stores only basic metadata. If not set, defaults to
+     * "everything".
+     */
+    data_storage_setting?: 'everything' | 'everything_except_pii' | 'basic_attributes_only' | null;
+    /**
+     * Body param: If users stay silent for a period after agent speech, end the chat.
+     * The minimum value allowed is 120,000 ms (2 minutes). The maximum value allowed
+     * is 259,200,000 ms (72 hours). By default, this is set to 3,600,000 (1 hour).
+     */
+    end_chat_after_silence_ms?: number | null;
+    /**
+     * Body param: Configuration for guardrail checks to detect and prevent prohibited
+     * topics in agent output and user input.
+     */
+    guardrail_config?: ChatAgentUpdateParams.GuardrailConfig;
+    /**
+     * Body param: Toggle behavior presets on/off to influence agent response style and
+     * behaviors. Voice-only presets are not available for chat agents.
+     */
+    handbook_config?: ChatAgentUpdateParams.HandbookConfig;
+    /**
+     * Body param: Specifies what language(s) the agent will operate in. Accepts either
+     * a single locale (e.g. `en-US`) or an array of locales for multilingual agents
+     * (e.g. `["en-US","es-ES"]`). The scalar value `multi` is deprecated but still
+     * accepted as a scalar, and is stored and returned as the ten locales it used to
+     * mean. It must not appear inside the array form. Send an explicit locale array
+     * instead. If unset, defaults to `en-US`.
+     */
+    language?: 'en-US' | 'en-IN' | 'en-GB' | 'en-AU' | 'en-NZ' | 'de-DE' | 'es-ES' | 'es-419' | 'hi-IN' | 'fr-FR' | 'fr-CA' | 'ja-JP' | 'pt-PT' | 'pt-BR' | 'zh-CN' | 'ru-RU' | 'it-IT' | 'ko-KR' | 'nl-NL' | 'nl-BE' | 'pl-PL' | 'tr-TR' | 'vi-VN' | 'ro-RO' | 'bg-BG' | 'ca-ES' | 'th-TH' | 'da-DK' | 'fi-FI' | 'el-GR' | 'hu-HU' | 'id-ID' | 'no-NO' | 'sk-SK' | 'sv-SE' | 'lt-LT' | 'lv-LV' | 'cs-CZ' | 'ms-MY' | 'af-ZA' | 'ar-SA' | 'az-AZ' | 'bs-BA' | 'cy-GB' | 'fa-IR' | 'fil-PH' | 'gl-ES' | 'he-IL' | 'hr-HR' | 'hy-AM' | 'is-IS' | 'kk-KZ' | 'kn-IN' | 'mk-MK' | 'mr-IN' | 'ne-NP' | 'sl-SI' | 'sr-RS' | 'sw-KE' | 'ta-IN' | 'ur-IN' | 'yue-CN' | 'uk-UA' | 'multi' | Array<'en-US' | 'en-IN' | 'en-GB' | 'en-AU' | 'en-NZ' | 'de-DE' | 'es-ES' | 'es-419' | 'hi-IN' | 'fr-FR' | 'fr-CA' | 'ja-JP' | 'pt-PT' | 'pt-BR' | 'zh-CN' | 'ru-RU' | 'it-IT' | 'ko-KR' | 'nl-NL' | 'nl-BE' | 'pl-PL' | 'tr-TR' | 'vi-VN' | 'ro-RO' | 'bg-BG' | 'ca-ES' | 'th-TH' | 'da-DK' | 'fi-FI' | 'el-GR' | 'hu-HU' | 'id-ID' | 'no-NO' | 'sk-SK' | 'sv-SE' | 'lt-LT' | 'lv-LV' | 'cs-CZ' | 'ms-MY' | 'af-ZA' | 'ar-SA' | 'az-AZ' | 'bs-BA' | 'cy-GB' | 'fa-IR' | 'fil-PH' | 'gl-ES' | 'he-IL' | 'hr-HR' | 'hy-AM' | 'is-IS' | 'kk-KZ' | 'kn-IN' | 'mk-MK' | 'mr-IN' | 'ne-NP' | 'sl-SI' | 'sr-RS' | 'sw-KE' | 'ta-IN' | 'ur-IN' | 'yue-CN' | 'uk-UA'>;
+    /**
+     * Body param: Whether this agent opts in to signed url for public log. If not set,
+     * default value of false will apply.
+     */
+    opt_in_signed_url?: boolean;
+    /**
+     * Body param: Configuration for PII scrubbing from transcripts and recordings.
+     */
+    pii_config?: ChatAgentUpdateParams.PiiConfig;
+    /**
+     * Body param: Post chat analysis data to extract from the chat. This data will
+     * augment the pre-defined variables extracted in the chat analysis. This will be
+     * available after the chat ends.
+     */
+    post_chat_analysis_data?: Array<ChatAgentUpdateParams.StringAnalysisData | ChatAgentUpdateParams.EnumAnalysisData | ChatAgentUpdateParams.BooleanAnalysisData | ChatAgentUpdateParams.NumberAnalysisData | ChatAgentUpdateParams.ChatPresetAnalysisData> | null;
+    /**
+     * Body param: The model to use for post chat analysis. Default to gpt-5.6-terra.
+     */
+    post_chat_analysis_model?: 'gpt-4.1' | 'gpt-4.1-mini' | 'gpt-4.1-nano' | 'gpt-5' | 'gpt-5-mini' | 'gpt-5-nano' | 'gpt-5.1' | 'gpt-5.2' | 'gpt-5.4' | 'gpt-5.4-mini' | 'gpt-5.4-nano' | 'gpt-5.5' | 'gpt-5.6-terra' | 'gpt-5.6-luna' | 'gpt-6-astra' | 'gpt-6-sol' | 'gpt-6.1-sol' | 'gpt-6-luna' | 'claude-4.5-sonnet' | 'claude-4.6-sonnet' | 'claude-5-opus' | 'claude-5.5-opus' | 'claude-5-sonnet' | 'claude-5.5-sonnet' | 'claude-4.5-haiku' | 'gemini-3.0-flash' | 'gemini-3.1-flash-lite' | 'gemini-3.5-flash' | 'gemini-3.5-flash-lite' | 'gemini-3.6-flash' | 'gemini-3.7-flash' | 'gemini-3.8-flash' | null;
+    /**
+     * Body param: Integration (Agent Functions) tools run as a dependency graph at
+     * chat end, after post-chat analysis. Each tool can be gated by a condition. Set
+     * to null to clear.
+     */
+    post_session_tools?: Array<ChatAgentUpdateParams.AppTool | ChatAgentUpdateParams.CustomTool | ChatAgentUpdateParams.CodeTool | ChatAgentUpdateParams.SendSMSTool> | null;
+    /**
+     * Body param: Integration (Agent Functions) tools run as a dependency graph before
+     * the chat's first message. Outputs are injected as dynamic variables. Set to null
+     * to clear.
+     */
+    pre_session_tools?: Array<ChatAgentUpdateParams.AppTool | ChatAgentUpdateParams.CustomTool | ChatAgentUpdateParams.CodeTool> | null;
+    /**
+     * Body param: The Response Engine to attach to the agent. It is used to generate
+     * responses for the agent. You need to create a Response Engine first before
+     * attaching it to an agent.
+     */
+    response_engine?: ChatAgentUpdateParams.ResponseEngineRetellLm | ChatAgentUpdateParams.ResponseEngineCustomLm | ChatAgentUpdateParams.ResponseEngineConversationFlow;
+    /**
+     * Body param: The expiration time for the signed url in milliseconds. Only
+     * applicable when opt_in_signed_url is true. If not set, default value of 86400000
+     * (24 hours) will apply.
+     */
+    signed_url_expiration_ms?: number | null;
+    /**
+     * Body param: IANA timezone for the agent (e.g. America/New_York). Defaults to
+     * America/Los_Angeles if not set.
+     */
+    timezone?: string | null;
+    /**
+     * Body param: Optional title of the chat agent version. Used for your own
+     * reference.
+     */
+    version_title?: string | null;
+    /**
+     * Body param: Which webhook events this agent should receive. If not set, defaults
+     * to chat_started, chat_ended, chat_analyzed.
+     */
+    webhook_events?: Array<'chat_started' | 'chat_ended' | 'chat_analyzed' | 'transcript_updated'> | null;
+    /**
+     * Body param: The timeout for the webhook in milliseconds. If not set, default
+     * value of 10000 will apply.
+     */
+    webhook_timeout_ms?: number;
+    /**
+     * Body param: The webhook for agent to listen to chat events. See what events it
+     * would get at [webhook doc](/features/webhook). If set, will binds webhook events
+     * for this agent to the specified url, and will ignore the account level webhook
+     * for this agent. Set to `null` to remove webhook url from this agent.
+     */
+    webhook_url?: string | null;
+}
+export declare namespace ChatAgentUpdateParams {
+    /**
+     * Contact memory settings for phone calls and SMS chats. Creating an agent
+     * defaults enable_update to false and enable_read to true. Updates only change the
+     * supplied flags; omitted flags stay unchanged and an empty object has no effect.
+     * Set a flag to false to disable it. The configuration cannot be cleared. Existing
+     * agents without this configuration have both disabled.
+     */
+    interface ContactMemoryConfig {
+        /**
+         * Automatically add saved contact memory to the agent prompt. Skippable nodes can
+         * use answers from the current conversation even when this setting is disabled.
+         * Contact dynamic variables, including contact_memory, remain available regardless
+         * of this setting.
+         */
+        enable_read?: boolean;
+        /**
+         * Rewrite the contact memory after each conversation. Requires storing
+         * conversation data. Chat agents must also have end_chat_after_silence_ms set.
+         */
+        enable_update?: boolean;
+    }
+    /**
+     * Configuration for guardrail checks to detect and prevent prohibited topics in
+     * agent output and user input.
+     */
+    interface GuardrailConfig {
+        /**
+         * Selected prohibited user topic categories to check. When user messages contain
+         * these topics, the agent will respond with a placeholder message instead of
+         * processing the request.
+         */
+        input_topics?: Array<'platform_integrity_jailbreaking'> | null;
+        /**
+         * Selected prohibited agent topic categories to check. When agent messages contain
+         * these topics, they will be replaced with a placeholder message.
+         */
+        output_topics?: Array<'harassment' | 'self_harm' | 'sexual_exploitation' | 'violence' | 'defense_and_national_security' | 'illicit_and_harmful_activity' | 'gambling' | 'regulated_professional_advice' | 'child_safety_and_exploitation'> | null;
+    }
+    /**
+     * Toggle behavior presets on/off to influence agent response style and behaviors.
+     * Voice-only presets are not available for chat agents.
+     */
+    interface HandbookConfig {
+        /**
+         * When asked, acknowledge being a virtual assistant.
+         */
+        ai_disclosure?: boolean;
+        /**
+         * Professional call center rep baseline.
+         */
+        default_personality?: boolean;
+        /**
+         * Warm acknowledgment of caller concerns.
+         */
+        high_empathy?: boolean;
+        /**
+         * Stay within prompt/context scope, don't invent details.
+         */
+        scope_boundaries?: boolean;
+    }
+    /**
+     * Configuration for PII scrubbing from transcripts and recordings.
+     */
+    interface PiiConfig {
+        /**
+         * List of PII categories to scrub from transcripts and recordings. PII redaction
+         * is only active when this list is non-empty; an empty array means no PII
+         * scrubbing is performed.
+         */
+        categories: Array<'person_name' | 'address' | 'email' | 'phone_number' | 'ssn' | 'passport' | 'driver_license' | 'credit_card' | 'bank_account' | 'password' | 'pin' | 'medical_id' | 'date_of_birth' | 'customer_account_number'>;
+        /**
+         * The processing mode for PII scrubbing. Currently only post-call is supported.
+         */
+        mode: 'post_call';
+    }
+    interface StringAnalysisData {
+        /**
+         * Description of the variable.
+         */
+        description: string;
+        /**
+         * Name of the variable.
+         */
+        name: string;
+        /**
+         * Type of the variable to extract.
+         */
+        type: 'string';
+        /**
+         * Optional instruction to help decide whether this field needs to be populated in
+         * the analysis. If not set, the field is always included. If required is true,
+         * this is ignored.
+         */
+        conditional_prompt?: string;
+        /**
+         * Examples of the variable value to teach model the style and syntax.
+         */
+        examples?: Array<string>;
+        /**
+         * Whether this data is required. If true and the data is not extracted, the call
+         * will be marked as unsuccessful.
+         */
+        required?: boolean;
+    }
+    interface EnumAnalysisData {
+        /**
+         * The possible values of the variable, must be non empty array.
          */
         choices: Array<string>;
         /**
@@ -840,316 +2972,856 @@ export declare namespace ChatAgentCreateParams {
          */
         required?: boolean;
     }
-}
-export interface ChatAgentRetrieveParams {
-    /**
-     * Optional version of the API to use for this request. If not provided, will
-     * default to latest version.
-     */
-    version?: string | number;
-}
-export interface ChatAgentUpdateParams {
-    /**
-     * Query param: Optional version of the API to use for this request. Default to
-     * latest version.
-     */
-    version?: string | number;
-    /**
-     * Body param: The name of the chat agent. Only used for your own reference.
-     */
-    agent_name?: string | null;
-    /**
-     * Body param: Message to display when the chat is automatically closed.
-     */
-    auto_close_message?: string | null;
-    /**
-     * Body param: Number of days to retain call/chat data before automatic deletion.
-     * Must be between 1 and 730 days. If not set, data is retained forever (no
-     * automatic deletion).
-     */
-    data_storage_retention_days?: number | null;
-    /**
-     * Body param: Controls what data is stored for this agent. "everything" stores all
-     * data including transcripts and recordings. "everything_except_pii" stores data
-     * but excludes PII when possible based on PII configuration.
-     * "basic_attributes_only" stores only basic metadata. If not set, defaults to
-     * "everything".
-     */
-    data_storage_setting?: 'everything' | 'everything_except_pii' | 'basic_attributes_only' | null;
-    /**
-     * Body param: If users stay silent for a period after agent speech, end the chat.
-     * The minimum value allowed is 120,000 ms (2 minutes). The maximum value allowed
-     * is 259,200,000 ms (72 hours). By default, this is set to 3,600,000 (1 hour).
-     */
-    end_chat_after_silence_ms?: number | null;
-    /**
-     * Body param: Configuration for guardrail checks to detect and prevent prohibited
-     * topics in agent output and user input.
-     */
-    guardrail_config?: ChatAgentUpdateParams.GuardrailConfig;
-    /**
-     * Body param: Toggle behavior presets on/off to influence agent response style and
-     * behaviors. Voice-only presets are not available for chat agents.
-     */
-    handbook_config?: ChatAgentUpdateParams.HandbookConfig;
-    /**
-     * Body param: Specifies what language(s) the agent will operate in. Accepts either
-     * a single locale (e.g. `en-US`) or an array of locales for multilingual agents
-     * (e.g. `["en-US","es-ES"]`). The scalar value `multi` is deprecated but still
-     * accepted as a scalar, and is stored and returned as the ten locales it used to
-     * mean. It must not appear inside the array form. Send an explicit locale array
-     * instead. If unset, defaults to `en-US`.
-     */
-    language?: 'en-US' | 'en-IN' | 'en-GB' | 'en-AU' | 'en-NZ' | 'de-DE' | 'es-ES' | 'es-419' | 'hi-IN' | 'fr-FR' | 'fr-CA' | 'ja-JP' | 'pt-PT' | 'pt-BR' | 'zh-CN' | 'ru-RU' | 'it-IT' | 'ko-KR' | 'nl-NL' | 'nl-BE' | 'pl-PL' | 'tr-TR' | 'vi-VN' | 'ro-RO' | 'bg-BG' | 'ca-ES' | 'th-TH' | 'da-DK' | 'fi-FI' | 'el-GR' | 'hu-HU' | 'id-ID' | 'no-NO' | 'sk-SK' | 'sv-SE' | 'lt-LT' | 'lv-LV' | 'cs-CZ' | 'ms-MY' | 'af-ZA' | 'ar-SA' | 'az-AZ' | 'bs-BA' | 'cy-GB' | 'fa-IR' | 'fil-PH' | 'gl-ES' | 'he-IL' | 'hr-HR' | 'hy-AM' | 'is-IS' | 'kk-KZ' | 'kn-IN' | 'mk-MK' | 'mr-IN' | 'ne-NP' | 'sl-SI' | 'sr-RS' | 'sw-KE' | 'ta-IN' | 'ur-IN' | 'yue-CN' | 'uk-UA' | 'multi' | Array<'en-US' | 'en-IN' | 'en-GB' | 'en-AU' | 'en-NZ' | 'de-DE' | 'es-ES' | 'es-419' | 'hi-IN' | 'fr-FR' | 'fr-CA' | 'ja-JP' | 'pt-PT' | 'pt-BR' | 'zh-CN' | 'ru-RU' | 'it-IT' | 'ko-KR' | 'nl-NL' | 'nl-BE' | 'pl-PL' | 'tr-TR' | 'vi-VN' | 'ro-RO' | 'bg-BG' | 'ca-ES' | 'th-TH' | 'da-DK' | 'fi-FI' | 'el-GR' | 'hu-HU' | 'id-ID' | 'no-NO' | 'sk-SK' | 'sv-SE' | 'lt-LT' | 'lv-LV' | 'cs-CZ' | 'ms-MY' | 'af-ZA' | 'ar-SA' | 'az-AZ' | 'bs-BA' | 'cy-GB' | 'fa-IR' | 'fil-PH' | 'gl-ES' | 'he-IL' | 'hr-HR' | 'hy-AM' | 'is-IS' | 'kk-KZ' | 'kn-IN' | 'mk-MK' | 'mr-IN' | 'ne-NP' | 'sl-SI' | 'sr-RS' | 'sw-KE' | 'ta-IN' | 'ur-IN' | 'yue-CN' | 'uk-UA'>;
-    /**
-     * Body param: Whether this agent opts in to signed url for public log. If not set,
-     * default value of false will apply.
-     */
-    opt_in_signed_url?: boolean;
-    /**
-     * Body param: Configuration for PII scrubbing from transcripts and recordings.
-     */
-    pii_config?: ChatAgentUpdateParams.PiiConfig;
-    /**
-     * Body param: Post chat analysis data to extract from the chat. This data will
-     * augment the pre-defined variables extracted in the chat analysis. This will be
-     * available after the chat ends.
-     */
-    post_chat_analysis_data?: Array<ChatAgentUpdateParams.StringAnalysisData | ChatAgentUpdateParams.EnumAnalysisData | ChatAgentUpdateParams.BooleanAnalysisData | ChatAgentUpdateParams.NumberAnalysisData | ChatAgentUpdateParams.ChatPresetAnalysisData> | null;
-    /**
-     * Body param: The model to use for post chat analysis. Default to gpt-5.6-terra.
-     */
-    post_chat_analysis_model?: 'gpt-4.1' | 'gpt-4.1-mini' | 'gpt-4.1-nano' | 'gpt-5' | 'gpt-5-mini' | 'gpt-5-nano' | 'gpt-5.1' | 'gpt-5.2' | 'gpt-5.4' | 'gpt-5.4-mini' | 'gpt-5.4-nano' | 'gpt-5.5' | 'gpt-5.6-terra' | 'gpt-5.6-luna' | 'claude-4.5-sonnet' | 'claude-4.6-sonnet' | 'claude-5-sonnet' | 'claude-4.5-haiku' | 'gemini-3.0-flash' | 'gemini-3.1-flash-lite' | 'gemini-3.5-flash' | 'gemini-3.5-flash-lite' | 'gemini-3.6-flash' | 'gemini-3.7-flash' | 'gemini-3.8-flash' | null;
-    /**
-     * Body param: The Response Engine to attach to the agent. It is used to generate
-     * responses for the agent. You need to create a Response Engine first before
-     * attaching it to an agent.
-     */
-    response_engine?: ChatAgentUpdateParams.ResponseEngineRetellLm | ChatAgentUpdateParams.ResponseEngineCustomLm | ChatAgentUpdateParams.ResponseEngineConversationFlow;
-    /**
-     * Body param: The expiration time for the signed url in milliseconds. Only
-     * applicable when opt_in_signed_url is true. If not set, default value of 86400000
-     * (24 hours) will apply.
-     */
-    signed_url_expiration_ms?: number | null;
-    /**
-     * Body param: IANA timezone for the agent (e.g. America/New_York). Defaults to
-     * America/Los_Angeles if not set.
-     */
-    timezone?: string | null;
-    /**
-     * Body param: Optional title of the chat agent version. Used for your own
-     * reference.
-     */
-    version_title?: string | null;
-    /**
-     * Body param: Which webhook events this agent should receive. If not set, defaults
-     * to chat_started, chat_ended, chat_analyzed.
-     */
-    webhook_events?: Array<'chat_started' | 'chat_ended' | 'chat_analyzed' | 'transcript_updated'> | null;
-    /**
-     * Body param: The timeout for the webhook in milliseconds. If not set, default
-     * value of 10000 will apply.
-     */
-    webhook_timeout_ms?: number;
-    /**
-     * Body param: The webhook for agent to listen to chat events. See what events it
-     * would get at [webhook doc](/features/webhook). If set, will binds webhook events
-     * for this agent to the specified url, and will ignore the account level webhook
-     * for this agent. Set to `null` to remove webhook url from this agent.
-     */
-    webhook_url?: string | null;
-}
-export declare namespace ChatAgentUpdateParams {
-    /**
-     * Configuration for guardrail checks to detect and prevent prohibited topics in
-     * agent output and user input.
-     */
-    interface GuardrailConfig {
+    interface AppTool {
+        /**
+         * The connection (App) this tool runs against. Must be a connection in the
+         * organization whose provider matches this tool's provider.
+         */
+        app_id: string;
+        /**
+         * Name of the catalog template within the provider, as listed by
+         * list-app-templates.
+         */
+        app_tool_template_name: string;
+        /**
+         * Name of the tool. Must be unique within the phase's tools; referenced by
+         * depends_on.
+         */
+        name: string;
+        /**
+         * Provider of the connection. Must match the connection's provider; supported
+         * providers are listed by list-app-templates.
+         */
+        provider: string;
+        type: 'integration_app';
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        condition?: AppTool.Condition;
+        /**
+         * Names of tools that must run before this one.
+         */
+        depends_on?: Array<string>;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Overrides the catalog template's LLM-facing description.
+         */
+        description?: string;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. If true, play a typing sound on the agent audio track while this
+         * tool is executing. Useful when the tool takes a noticeable amount of time to
+         * prevent silence on the call.
+         */
+        enable_typing_sound?: boolean;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. The message for the agent to speak when executing the tool. Only
+         * applicable when speak_during_execution is true.
+         */
+        execution_message_description?: string;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * What the agent and the transcript see of the tool's response. Omit to send the
+         * full response. Does not affect response_variables, which are always extracted
+         * from the raw response.
+         */
+        output_selection?: AppTool.UnionMember0 | AppTool.UnionMember1;
+        /**
+         * The resolved input parameters, in order. Properties may pin a value with const
+         * (including {{variable}} references) or provide a description for LLM inference.
+         * Each property may also record selected*input_mode, the editor mode the user
+         * selected ("const_enum", "const_boolean", "const_value", "description_custom", or
+         * "description_preset"); it is stored and returned as-is, used only by the tool
+         * config UI. Omit the key when no mode is recorded; when set, const*_ modes
+         * require a non-empty const, and description\__ modes must omit const entirely.
+         * Each parameter's required list must match the schema returned by the
+         * corresponding step of the get-app-tool-schema loop.
+         */
+        parameters?: Array<AppTool.Parameter>;
+        /**
+         * Mapping of a dynamic-variable name to the response field (dot-path) it is
+         * populated from. Missing paths are ignored.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Determines whether the agent would call LLM another time and speak
+         * when the result of the tool is obtained.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. If true, will speak during execution.
+         */
+        speak_during_execution?: boolean;
+    }
+    namespace AppTool {
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        interface Condition {
+            equations: Array<Condition.Equation>;
+            operator: '||' | '&&';
+            type: 'equation';
+        }
+        namespace Condition {
+            interface Equation {
+                /**
+                 * Left side of the equation
+                 */
+                left: string;
+                operator: '==' | '!=' | '>' | '>=' | '<' | '<=' | 'contains' | 'not_contains' | 'exists' | 'not_exist';
+                /**
+                 * Right side of the equation. The right side of the equation not required when
+                 * "exists" or "not_exist" are selected.
+                 */
+                right?: string;
+            }
+        }
+        interface UnionMember0 {
+            mode: 'all';
+            /**
+             * Not used at runtime; stored and returned as-is for the UI.
+             */
+            fields?: Array<string>;
+        }
+        interface UnionMember1 {
+            /**
+             * The only response fields the agent and the transcript see, as dot-paths into the
+             * response schema returned by get-app-tool-schema. Everything else is dropped.
+             * Selecting a parent keeps its whole subtree. A plain segment traverses arrays
+             * element-wise (deals.properties.amount keeps that field on every deal), while
+             * key[n] selects one element (deals[0].id keeps only the first deal's id); paths
+             * that match nothing contribute nothing.
+             */
+            fields: Array<string>;
+            mode: 'subset';
+        }
+        /**
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
+         */
+        interface Parameter {
+            /**
+             * The value of properties is an object, where each key is the name of a property
+             * and each value is a schema used to validate that property.
+             */
+            properties: unknown;
+            /**
+             * Type must be "object" for a JSON Schema object.
+             */
+            type: 'object';
+            /**
+             * List of names of required property when generating this parameter. LLM will do
+             * its best to generate the required properties in its function arguments. Property
+             * must exist in properties.
+             */
+            required?: Array<string>;
+        }
+    }
+    interface CustomTool {
+        /**
+         * Name of the tool. Must be unique within all tools available to LLM at any given
+         * time (general tools + state tools + state edges). Must be consisted of a-z, A-Z,
+         * 0-9, or contain underscores and dashes, with a maximum length of 64 (no space
+         * allowed).
+         */
+        name: string;
+        type: 'custom';
+        /**
+         * Describes what the tool does, sometimes can also include information about when
+         * to call the tool.
+         */
+        url: string;
+        /**
+         * If set to true, the parameters will be passed as root level JSON object instead
+         * of nested under "args".
+         */
+        args_at_root?: boolean;
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        condition?: CustomTool.Condition;
+        /**
+         * Names of tools that must run before this one.
+         */
+        depends_on?: Array<string>;
+        /**
+         * Describes what this tool does and when to call this tool.
+         */
+        description?: string;
+        /**
+         * If true, play a typing sound on the agent audio track while this tool is
+         * executing. Useful when the tool takes a noticeable amount of time to prevent
+         * silence on the call.
+         */
+        enable_typing_sound?: boolean;
         /**
-         * Selected prohibited user topic categories to check. When user messages contain
-         * these topics, the agent will respond with a placeholder message instead of
-         * processing the request.
+         * The description for the sentence agent say during execution. Only applicable
+         * when speak_during_execution is true. Can write what to say or even provide
+         * examples. The default is "The message you will say to callee when calling this
+         * tool. Make sure it fits into the conversation smoothly.".
          */
-        input_topics?: Array<'platform_integrity_jailbreaking'> | null;
+        execution_message_description?: string;
         /**
-         * Selected prohibited agent topic categories to check. When agent messages contain
-         * these topics, they will be replaced with a placeholder message.
+         * Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
          */
-        output_topics?: Array<'harassment' | 'self_harm' | 'sexual_exploitation' | 'violence' | 'defense_and_national_security' | 'illicit_and_harmful_activity' | 'gambling' | 'regulated_professional_advice' | 'child_safety_and_exploitation'> | null;
-    }
-    /**
-     * Toggle behavior presets on/off to influence agent response style and behaviors.
-     * Voice-only presets are not available for chat agents.
-     */
-    interface HandbookConfig {
+        execution_message_type?: 'prompt' | 'static_text';
         /**
-         * When asked, acknowledge being a virtual assistant.
+         * Headers to add to the request.
          */
-        ai_disclosure?: boolean;
+        headers?: {
+            [key: string]: string;
+        };
         /**
-         * Professional call center rep baseline.
+         * Maximum number of times to retry the request after a failed attempt, from 0 (no
+         * retry) to 5. Retries happen on any failure, with exponential backoff between
+         * attempts; the backoff delay is not configurable. `timeout_ms` applies per
+         * attempt rather than as a budget across all attempts, so an attempt that times
+         * out is still retried and the worst-case total duration is `timeout_ms`
+         * multiplied by (`max_retry` + 1) as well as any latency incurred by the
+         * exponential backoff + jitter between each retry. Only the final attempt's result
+         * is reported to the agent. Because retries repeat the request, only set this
+         * above 0 if your endpoint is idempotent — a retried request may be processed more
+         * than once. Defaults to 0 (no retry).
          */
-        default_personality?: boolean;
+        max_retry?: number;
         /**
-         * Warm acknowledgment of caller concerns.
+         * Method to use for the request, default to POST.
          */
-        high_empathy?: boolean;
+        method?: 'GET' | 'POST' | 'PUT' | 'PATCH' | 'DELETE';
         /**
-         * Stay within prompt/context scope, don't invent details.
+         * How the tool's `parameters` are authored and shown in the dashboard editor —
+         * "form" for the visual parameter builder, "json" for a raw JSON Schema. Both
+         * produce the same `parameters` schema; this does not change how the request body
+         * is encoded (see `args_at_root`).
          */
-        scope_boundaries?: boolean;
+        parameter_type?: 'json' | 'form';
+        /**
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
+         */
+        parameters?: CustomTool.Parameters;
+        /**
+         * Query parameters to append to the request URL.
+         */
+        query_params?: {
+            [key: string]: string;
+        };
+        /**
+         * A mapping of variable names to JSON paths in the response body. These values
+         * will be extracted from the response and made available as dynamic variables for
+         * use.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Determines whether the agent would call LLM another time and speak when the
+         * result of function is obtained. Usually this needs to get turned on so user can
+         * get update for the function call.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Determines whether the agent would say sentence like "One moment, let me check
+         * that." when executing the function. Recommend to turn on if your function call
+         * takes over 1s (including network) to complete, so that your agent remains
+         * responsive.
+         */
+        speak_during_execution?: boolean;
+        /**
+         * The maximum time in milliseconds the tool can run before it's considered
+         * timeout. If the tool times out, the agent would have that info. The minimum
+         * value allowed is 1000 ms (1 s), and maximum value allowed is 600,000 ms (10
+         * min). By default, this is set to 120,000 ms (2 min).
+         */
+        timeout_ms?: number;
     }
-    /**
-     * Configuration for PII scrubbing from transcripts and recordings.
-     */
-    interface PiiConfig {
+    namespace CustomTool {
         /**
-         * List of PII categories to scrub from transcripts and recordings. PII redaction
-         * is only active when this list is non-empty; an empty array means no PII
-         * scrubbing is performed.
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
          */
-        categories: Array<'person_name' | 'address' | 'email' | 'phone_number' | 'ssn' | 'passport' | 'driver_license' | 'credit_card' | 'bank_account' | 'password' | 'pin' | 'medical_id' | 'date_of_birth' | 'customer_account_number'>;
+        interface Condition {
+            equations: Array<Condition.Equation>;
+            operator: '||' | '&&';
+            type: 'equation';
+        }
+        namespace Condition {
+            interface Equation {
+                /**
+                 * Left side of the equation
+                 */
+                left: string;
+                operator: '==' | '!=' | '>' | '>=' | '<' | '<=' | 'contains' | 'not_contains' | 'exists' | 'not_exist';
+                /**
+                 * Right side of the equation. The right side of the equation not required when
+                 * "exists" or "not_exist" are selected.
+                 */
+                right?: string;
+            }
+        }
         /**
-         * The processing mode for PII scrubbing. Currently only post-call is supported.
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
          */
-        mode: 'post_call';
+        interface Parameters {
+            /**
+             * The value of properties is an object, where each key is the name of a property
+             * and each value is a schema used to validate that property.
+             */
+            properties: unknown;
+            /**
+             * Type must be "object" for a JSON Schema object.
+             */
+            type: 'object';
+            /**
+             * List of names of required property when generating this parameter. LLM will do
+             * its best to generate the required properties in its function arguments. Property
+             * must exist in properties.
+             */
+            required?: Array<string>;
+        }
     }
-    interface StringAnalysisData {
+    interface CodeTool {
         /**
-         * Description of the variable.
+         * JavaScript code to execute in the sandbox.
          */
-        description: string;
+        code: string;
         /**
-         * Name of the variable.
+         * Name of the tool. Must be unique within all tools available to LLM at any given
+         * time (general tools + state tools + state edges). Must be consisted of a-z, A-Z,
+         * 0-9, or contain underscores and dashes, with a maximum length of 64 (no space
+         * allowed).
          */
         name: string;
+        type: 'code';
         /**
-         * Type of the variable to extract.
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
          */
-        type: 'string';
+        condition?: CodeTool.Condition;
         /**
-         * Optional instruction to help decide whether this field needs to be populated in
-         * the analysis. If not set, the field is always included. If required is true,
-         * this is ignored.
+         * Names of tools that must run before this one.
          */
-        conditional_prompt?: string;
+        depends_on?: Array<string>;
         /**
-         * Examples of the variable value to teach model the style and syntax.
+         * Describes what this tool does and when to call this tool.
          */
-        examples?: Array<string>;
+        description?: string;
         /**
-         * Whether this data is required. If true and the data is not extracted, the call
-         * will be marked as unsuccessful.
+         * If true, play a typing sound on the agent audio track while this tool is
+         * executing.
          */
-        required?: boolean;
-    }
-    interface EnumAnalysisData {
+        enable_typing_sound?: boolean;
         /**
-         * The possible values of the variable, must be non empty array.
+         * The description for the sentence agent say during execution. Only applicable
+         * when speak_during_execution is true.
          */
-        choices: Array<string>;
+        execution_message_description?: string;
         /**
-         * Description of the variable.
+         * Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
          */
-        description: string;
+        execution_message_type?: 'prompt' | 'static_text';
         /**
-         * Name of the variable.
+         * A mapping of variable names to JSON paths in the code execution result. These
+         * mapped values will be extracted and added as dynamic variables.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Determines whether the agent would call LLM another time and speak when the
+         * result of function is obtained.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Determines whether the agent would say sentence like "One moment, let me check
+         * that." when executing the tool.
+         */
+        speak_during_execution?: boolean;
+        /**
+         * The maximum time in milliseconds the code can run before it's considered
+         * timeout. Defaults to 30,000 ms (30 s).
+         */
+        timeout_ms?: number;
+    }
+    namespace CodeTool {
+        /**
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
+         */
+        interface Condition {
+            equations: Array<Condition.Equation>;
+            operator: '||' | '&&';
+            type: 'equation';
+        }
+        namespace Condition {
+            interface Equation {
+                /**
+                 * Left side of the equation
+                 */
+                left: string;
+                operator: '==' | '!=' | '>' | '>=' | '<' | '<=' | 'contains' | 'not_contains' | 'exists' | 'not_exist';
+                /**
+                 * Right side of the equation. The right side of the equation not required when
+                 * "exists" or "not_exist" are selected.
+                 */
+                right?: string;
+            }
+        }
+    }
+    interface SendSMSTool {
+        /**
+         * Name of the tool. Must be unique within all tools available to LLM at any given
+         * time (general tools + state tools + state edges).
          */
         name: string;
+        sms_content: SendSMSTool.SMSContentPredefined | SendSMSTool.SMSContentInferred | SendSMSTool.SMSContentTemplate;
+        type: 'send_sms';
         /**
-         * Type of the variable to extract.
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
          */
-        type: 'enum';
+        condition?: SendSMSTool.Condition;
         /**
-         * Optional instruction to help decide whether this field needs to be populated in
-         * the analysis. If not set, the field is always included. If required is true,
-         * this is ignored.
+         * Names of tools that must run before this one.
          */
-        conditional_prompt?: string;
+        depends_on?: Array<string>;
         /**
-         * Whether this data is required. If true and the data is not extracted, the call
-         * will be marked as unsuccessful.
+         * Describes what the tool does, sometimes can also include information about when
+         * to call the tool.
          */
-        required?: boolean;
+        description?: string;
+        /**
+         * Describes what to say before sending the SMS. Only applicable when
+         * speak_during_execution is true.
+         */
+        execution_message_description?: string;
+        /**
+         * Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * If true, the agent will speak a short line before sending the SMS. If omitted,
+         * defaults to true (same as end_call / transfer_call tools).
+         */
+        speak_during_execution?: boolean;
     }
-    interface BooleanAnalysisData {
+    namespace SendSMSTool {
+        interface SMSContentPredefined {
+            /**
+             * The static message to be sent in the SMS. Can contain dynamic variables.
+             */
+            text?: string;
+            type?: 'predefined';
+        }
+        interface SMSContentInferred {
+            /**
+             * The prompt to be used to help infer the SMS content. The model will take the
+             * global prompt, the call transcript, and this prompt together to deduce the right
+             * message to send. Can contain dynamic variables.
+             */
+            prompt?: string;
+            type?: 'inferred';
+        }
+        interface SMSContentTemplate {
+            /**
+             * The template to use for the SMS content. "info_collection" sends a predefined
+             * message requesting information from the user.
+             */
+            template: 'info_collection';
+            type: 'template';
+        }
         /**
-         * Description of the variable.
+         * Optional gate; the step only runs when the condition holds. Defaults to always
+         * running.
          */
-        description: string;
+        interface Condition {
+            equations: Array<Condition.Equation>;
+            operator: '||' | '&&';
+            type: 'equation';
+        }
+        namespace Condition {
+            interface Equation {
+                /**
+                 * Left side of the equation
+                 */
+                left: string;
+                operator: '==' | '!=' | '>' | '>=' | '<' | '<=' | 'contains' | 'not_contains' | 'exists' | 'not_exist';
+                /**
+                 * Right side of the equation. The right side of the equation not required when
+                 * "exists" or "not_exist" are selected.
+                 */
+                right?: string;
+            }
+        }
+    }
+    interface AppTool {
         /**
-         * Name of the variable.
+         * The connection (App) this tool runs against. Must be a connection in the
+         * organization whose provider matches this tool's provider.
+         */
+        app_id: string;
+        /**
+         * Name of the catalog template within the provider, as listed by
+         * list-app-templates.
+         */
+        app_tool_template_name: string;
+        /**
+         * Name of the tool. Must be unique within the phase's tools; referenced by
+         * depends_on.
          */
         name: string;
         /**
-         * Type of the variable to extract.
+         * Provider of the connection. Must match the connection's provider; supported
+         * providers are listed by list-app-templates.
          */
-        type: 'boolean';
+        provider: string;
+        type: 'integration_app';
         /**
-         * Optional instruction to help decide whether this field needs to be populated in
-         * the analysis. If not set, the field is always included. If required is true,
-         * this is ignored.
+         * Names of tools that must run before this one.
          */
-        conditional_prompt?: string;
+        depends_on?: Array<string>;
         /**
-         * Whether this data is required. If true and the data is not extracted, the call
-         * will be marked as unsuccessful.
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Overrides the catalog template's LLM-facing description.
          */
-        required?: boolean;
+        description?: string;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. If true, play a typing sound on the agent audio track while this
+         * tool is executing. Useful when the tool takes a noticeable amount of time to
+         * prevent silence on the call.
+         */
+        enable_typing_sound?: boolean;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. The message for the agent to speak when executing the tool. Only
+         * applicable when speak_during_execution is true.
+         */
+        execution_message_description?: string;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * What the agent and the transcript see of the tool's response. Omit to send the
+         * full response. Does not affect response_variables, which are always extracted
+         * from the raw response.
+         */
+        output_selection?: AppTool.UnionMember0 | AppTool.UnionMember1;
+        /**
+         * The resolved input parameters, in order. Properties may pin a value with const
+         * (including {{variable}} references) or provide a description for LLM inference.
+         * Each property may also record selected*input_mode, the editor mode the user
+         * selected ("const_enum", "const_boolean", "const_value", "description_custom", or
+         * "description_preset"); it is stored and returned as-is, used only by the tool
+         * config UI. Omit the key when no mode is recorded; when set, const*_ modes
+         * require a non-empty const, and description\__ modes must omit const entirely.
+         * Each parameter's required list must match the schema returned by the
+         * corresponding step of the get-app-tool-schema loop.
+         */
+        parameters?: Array<AppTool.Parameter>;
+        /**
+         * Mapping of a dynamic-variable name to the response field (dot-path) it is
+         * populated from. Missing paths are ignored.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. Determines whether the agent would call LLM another time and speak
+         * when the result of the tool is obtained.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Only applies to during conversation functions; ignored by the pre/post
+         * conversation. If true, will speak during execution.
+         */
+        speak_during_execution?: boolean;
     }
-    interface NumberAnalysisData {
+    namespace AppTool {
+        interface UnionMember0 {
+            mode: 'all';
+            /**
+             * Not used at runtime; stored and returned as-is for the UI.
+             */
+            fields?: Array<string>;
+        }
+        interface UnionMember1 {
+            /**
+             * The only response fields the agent and the transcript see, as dot-paths into the
+             * response schema returned by get-app-tool-schema. Everything else is dropped.
+             * Selecting a parent keeps its whole subtree. A plain segment traverses arrays
+             * element-wise (deals.properties.amount keeps that field on every deal), while
+             * key[n] selects one element (deals[0].id keeps only the first deal's id); paths
+             * that match nothing contribute nothing.
+             */
+            fields: Array<string>;
+            mode: 'subset';
+        }
         /**
-         * Description of the variable.
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
          */
-        description: string;
+        interface Parameter {
+            /**
+             * The value of properties is an object, where each key is the name of a property
+             * and each value is a schema used to validate that property.
+             */
+            properties: unknown;
+            /**
+             * Type must be "object" for a JSON Schema object.
+             */
+            type: 'object';
+            /**
+             * List of names of required property when generating this parameter. LLM will do
+             * its best to generate the required properties in its function arguments. Property
+             * must exist in properties.
+             */
+            required?: Array<string>;
+        }
+    }
+    interface CustomTool {
         /**
-         * Name of the variable.
+         * Name of the tool. Must be unique within all tools available to LLM at any given
+         * time (general tools + state tools + state edges). Must be consisted of a-z, A-Z,
+         * 0-9, or contain underscores and dashes, with a maximum length of 64 (no space
+         * allowed).
          */
         name: string;
+        type: 'custom';
         /**
-         * Type of the variable to extract.
+         * Describes what the tool does, sometimes can also include information about when
+         * to call the tool.
          */
-        type: 'number';
+        url: string;
         /**
-         * Optional instruction to help decide whether this field needs to be populated in
-         * the analysis. If not set, the field is always included. If required is true,
-         * this is ignored.
+         * If set to true, the parameters will be passed as root level JSON object instead
+         * of nested under "args".
          */
-        conditional_prompt?: string;
+        args_at_root?: boolean;
         /**
-         * Whether this data is required. If true and the data is not extracted, the call
-         * will be marked as unsuccessful.
+         * Names of tools that must run before this one.
          */
-        required?: boolean;
+        depends_on?: Array<string>;
+        /**
+         * Describes what this tool does and when to call this tool.
+         */
+        description?: string;
+        /**
+         * If true, play a typing sound on the agent audio track while this tool is
+         * executing. Useful when the tool takes a noticeable amount of time to prevent
+         * silence on the call.
+         */
+        enable_typing_sound?: boolean;
+        /**
+         * The description for the sentence agent say during execution. Only applicable
+         * when speak_during_execution is true. Can write what to say or even provide
+         * examples. The default is "The message you will say to callee when calling this
+         * tool. Make sure it fits into the conversation smoothly.".
+         */
+        execution_message_description?: string;
+        /**
+         * Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * Headers to add to the request.
+         */
+        headers?: {
+            [key: string]: string;
+        };
+        /**
+         * Maximum number of times to retry the request after a failed attempt, from 0 (no
+         * retry) to 5. Retries happen on any failure, with exponential backoff between
+         * attempts; the backoff delay is not configurable. `timeout_ms` applies per
+         * attempt rather than as a budget across all attempts, so an attempt that times
+         * out is still retried and the worst-case total duration is `timeout_ms`
+         * multiplied by (`max_retry` + 1) as well as any latency incurred by the
+         * exponential backoff + jitter between each retry. Only the final attempt's result
+         * is reported to the agent. Because retries repeat the request, only set this
+         * above 0 if your endpoint is idempotent — a retried request may be processed more
+         * than once. Defaults to 0 (no retry).
+         */
+        max_retry?: number;
+        /**
+         * Method to use for the request, default to POST.
+         */
+        method?: 'GET' | 'POST' | 'PUT' | 'PATCH' | 'DELETE';
+        /**
+         * How the tool's `parameters` are authored and shown in the dashboard editor —
+         * "form" for the visual parameter builder, "json" for a raw JSON Schema. Both
+         * produce the same `parameters` schema; this does not change how the request body
+         * is encoded (see `args_at_root`).
+         */
+        parameter_type?: 'json' | 'form';
+        /**
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
+         */
+        parameters?: CustomTool.Parameters;
+        /**
+         * Query parameters to append to the request URL.
+         */
+        query_params?: {
+            [key: string]: string;
+        };
+        /**
+         * A mapping of variable names to JSON paths in the response body. These values
+         * will be extracted from the response and made available as dynamic variables for
+         * use.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Determines whether the agent would call LLM another time and speak when the
+         * result of function is obtained. Usually this needs to get turned on so user can
+         * get update for the function call.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Determines whether the agent would say sentence like "One moment, let me check
+         * that." when executing the function. Recommend to turn on if your function call
+         * takes over 1s (including network) to complete, so that your agent remains
+         * responsive.
+         */
+        speak_during_execution?: boolean;
+        /**
+         * The maximum time in milliseconds the tool can run before it's considered
+         * timeout. If the tool times out, the agent would have that info. The minimum
+         * value allowed is 1000 ms (1 s), and maximum value allowed is 600,000 ms (10
+         * min). By default, this is set to 120,000 ms (2 min).
+         */
+        timeout_ms?: number;
     }
-    /**
-     * System preset for post-chat analysis (chat agents). Use in
-     * post_chat_analysis_data to override prompts or mark fields optional.
-     */
-    interface ChatPresetAnalysisData {
+    namespace CustomTool {
         /**
-         * Preset identifier for chat agent analysis.
+         * The parameters the functions accepts, described as a JSON Schema object. See
+         * [JSON Schema reference](https://json-schema.org/understanding-json-schema/) for
+         * documentation about the format. Omitting parameters defines a function with an
+         * empty parameter list.
          */
-        name: 'chat_summary' | 'chat_successful' | 'user_sentiment';
+        interface Parameters {
+            /**
+             * The value of properties is an object, where each key is the name of a property
+             * and each value is a schema used to validate that property.
+             */
+            properties: unknown;
+            /**
+             * Type must be "object" for a JSON Schema object.
+             */
+            type: 'object';
+            /**
+             * List of names of required property when generating this parameter. LLM will do
+             * its best to generate the required properties in its function arguments. Property
+             * must exist in properties.
+             */
+            required?: Array<string>;
+        }
+    }
+    interface CodeTool {
         /**
-         * Identifies this item as a system preset.
+         * JavaScript code to execute in the sandbox.
          */
-        type: 'system-presets';
+        code: string;
         /**
-         * Optional instruction to help decide whether this field needs to be populated. If
-         * not set, the field is always included.
+         * Name of the tool. Must be unique within all tools available to LLM at any given
+         * time (general tools + state tools + state edges). Must be consisted of a-z, A-Z,
+         * 0-9, or contain underscores and dashes, with a maximum length of 64 (no space
+         * allowed).
          */
-        conditional_prompt?: string;
+        name: string;
+        type: 'code';
         /**
-         * Prompt or description for this preset.
+         * Names of tools that must run before this one.
+         */
+        depends_on?: Array<string>;
+        /**
+         * Describes what this tool does and when to call this tool.
          */
         description?: string;
         /**
-         * If false, this field is optional in the analysis. If true or unset, the field is
-         * required.
+         * If true, play a typing sound on the agent audio track while this tool is
+         * executing.
          */
-        required?: boolean;
+        enable_typing_sound?: boolean;
+        /**
+         * The description for the sentence agent say during execution. Only applicable
+         * when speak_during_execution is true.
+         */
+        execution_message_description?: string;
+        /**
+         * Type of execution message. "prompt" means the agent will use
+         * execution_message_description as a prompt to generate the message. "static_text"
+         * means the agent will speak the execution_message_description directly. Defaults
+         * to "prompt".
+         */
+        execution_message_type?: 'prompt' | 'static_text';
+        /**
+         * A mapping of variable names to JSON paths in the code execution result. These
+         * mapped values will be extracted and added as dynamic variables.
+         */
+        response_variables?: {
+            [key: string]: string;
+        };
+        /**
+         * Determines whether the agent would call LLM another time and speak when the
+         * result of function is obtained.
+         */
+        speak_after_execution?: boolean;
+        /**
+         * Determines whether the agent would say sentence like "One moment, let me check
+         * that." when executing the tool.
+         */
+        speak_during_execution?: boolean;
+        /**
+         * The maximum time in milliseconds the code can run before it's considered
+         * timeout. Defaults to 30,000 ms (30 s).
+         */
+        timeout_ms?: number;
     }
     interface ResponseEngineRetellLm {
         /**