retell-sdk 0.1.0-alpha.1

package/src/resources/llm.ts

@@ -0,0 +1,1643 @@
+// File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+import * as Core from "../core";
+import { APIResource } from "../resource";
+import * as LlmAPI from "./llm";
+
+export class Llm extends APIResource {
+  /**
+   * Create a new Retell LLM
+   */
+  create(body: LlmCreateParams, options?: Core.RequestOptions): Core.APIPromise<LlmResponse> {
+    return this._client.post('/create-retell-llm', { body, ...options });
+  }
+
+  /**
+   * Retrieve details of a specific Retell LLM
+   */
+  retrieve(llmId: string, options?: Core.RequestOptions): Core.APIPromise<LlmResponse> {
+    return this._client.get(`/get-retell-llm/${llmId}`, options);
+  }
+
+  /**
+   * Update an existing Retell LLM
+   */
+  update(llmId: string, body: LlmUpdateParams, options?: Core.RequestOptions): Core.APIPromise<LlmResponse> {
+    return this._client.patch(`/update-retell-llm/${llmId}`, { body, ...options });
+  }
+
+  /**
+   * List all retell LLM
+   */
+  list(options?: Core.RequestOptions): Core.APIPromise<LlmListResponse> {
+    return this._client.get('/list-retell-llms', options);
+  }
+
+  /**
+   * Delete an existing Retell LLM
+   */
+  delete(llmId: string, options?: Core.RequestOptions): Core.APIPromise<void> {
+    return this._client.delete(`/delete-retell-llm/${llmId}`, {
+      ...options,
+      headers: { Accept: '*/*', ...options?.headers },
+    });
+  }
+}
+
+export interface LlmResponse {
+  /**
+   * Last modification timestamp (milliseconds since epoch). Either the time of last
+   * update or creation if no updates available.
+   */
+  last_modification_timestamp: number;
+
+  /**
+   * Unique id of Retell LLM.
+   */
+  llm_id: string;
+
+  /**
+   * The LLM Websocket URL constructed from unique id of Retell LLM. Used in agent
+   * API to create / update agent.
+   */
+  llm_websocket_url: string;
+
+  /**
+   * First utterance said by the agent in the call. If not set, LLM will dynamically
+   * generate a message. If set to "", agent will wait for user to speak first.
+   */
+  begin_message?: string;
+
+  /**
+   * General prompt that's appended to system prompt no matter what state the agent
+   * is in.
+   *
+   * - System prompt (with state) = general prompt + state prompt.
+   *
+   * - System prompt (no state) = general prompt.
+   */
+  general_prompt?: string;
+
+  /**
+   * A list of tools the model may call (to get external knowledge, call API, etc).
+   * You can select from some common predefined tools like end call, transfer call,
+   * etc; or you can create your own custom tool (last option) for the LLM to use.
+   *
+   * - Tools of LLM (with state) = general tools + state tools + state transitions
+   *
+   * - Tools of LLM (no state) = general tools
+   */
+  general_tools?: Array<
+    | LlmResponse.EndCallTool
+    | LlmResponse.TransferCallTool
+    | LlmResponse.CheckAvailabilityCalTool
+    | LlmResponse.BookAppointmentCalTool
+    | LlmResponse.CustomTool
+  >;
+
+  /**
+   * Name of the starting state. Required if states is not empty.
+   */
+  starting_state?: string;
+
+  /**
+   * States of the LLM. This is to help reduce prompt length and tool choices when
+   * the call can be broken into distinct states. With shorter prompts and less
+   * tools, the LLM can better focus and follow the rules, minimizing hallucination.
+   * If this field is not set, the agent would only have general prompt and general
+   * tools (essentially one state).
+   */
+  states?: Array<LlmResponse.State>;
+}
+
+export namespace LlmResponse {
+  export interface EndCallTool {
+    /**
+     * Name of the tool. Must be unique within all tools available to LLM at any given
+     * time (general tools + state tools + state transitions).
+     */
+    name: string;
+
+    type: 'end_call';
+
+    /**
+     * Describes when to end the call.
+     */
+    description?: string;
+  }
+
+  export interface TransferCallTool {
+    /**
+     * 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;
+
+    /**
+     * The number to transfer to in E.164 format (a + and country code, then the phone
+     * number with no space or other special characters). For example, +16175551212.
+     */
+    number: string;
+
+    type: 'transfer_call';
+
+    /**
+     * Describes when to transfer the call.
+     */
+    description?: string;
+  }
+
+  export interface CheckAvailabilityCalTool {
+    /**
+     * Cal.com Api key that have access to the cal.com event you want to check
+     * availability for.
+     */
+    cal_api_key: string;
+
+    /**
+     * Cal.com event type id number for the cal.com event you want to check
+     * availability for.
+     */
+    event_type_id: number;
+
+    /**
+     * Name of the tool. Must be unique within all tools available to LLM at any given
+     * time (general tools + state tools + state transitions).
+     */
+    name: string;
+
+    type: 'check_availability_cal';
+
+    /**
+     * Describes when to check availability.
+     */
+    description?: string;
+
+    /**
+     * Timezone to be used when checking availability, must be in
+     * [IANA timezone database](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones).
+     * If not specified, will check if user specified timezone in call, and if not,
+     * will use the timezone of the Retell servers.
+     */
+    timezone?: string;
+  }
+
+  export interface BookAppointmentCalTool {
+    /**
+     * Cal.com Api key that have access to the cal.com event you want to book
+     * appointment.
+     */
+    cal_api_key: string;
+
+    /**
+     * Cal.com event type id number for the cal.com event you want to book appointment.
+     */
+    event_type_id: number;
+
+    /**
+     * Name of the tool. Must be unique within all tools available to LLM at any given
+     * time (general tools + state tools + state transitions).
+     */
+    name: string;
+
+    type: 'book_appointment_cal';
+
+    /**
+     * Describes when to book the appointment.
+     */
+    description?: string;
+
+    /**
+     * Timezone to be used when booking appointment, must be in
+     * [IANA timezone database](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones).
+     * If not specified, will check if user specified timezone in call, and if not,
+     * will use the timezone of the Retell servers.
+     */
+    timezone?: string;
+  }
+
+  export interface CustomTool {
+    /**
+     * Describes what this tool does and when to call this tool.
+     */
+    description: string;
+
+    /**
+     * 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;
+
+    /**
+     * 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;
+
+    type: 'custom';
+
+    /**
+     * The URL we will post the function name and arguments to get a result for the
+     * function. Usually this is your server.
+     */
+    url: string;
+
+    /**
+     * 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;
+
+    /**
+     * 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;
+  }
+
+  export 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.
+     */
+    export 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: Record<string, 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>;
+    }
+  }
+
+  export interface State {
+    /**
+     * Name of the state, must be unique for each state.
+     */
+    name: string;
+
+    /**
+     * Edges of the state define how and what state can be reached from this state.
+     */
+    edges?: Array<State.Edge>;
+
+    /**
+     * Prompt of the state, will be appended to the system prompt of LLM.
+     *
+     * - System prompt = general prompt + state prompt.
+     */
+    state_prompt?: string;
+
+    /**
+     * A list of tools specific to this state the model may call (to get external
+     * knowledge, call API, etc). You can select from some common predefined tools like
+     * end call, transfer call, etc; or you can create your own custom tool (last
+     * option) for the LLM to use.
+     *
+     * - Tools of LLM = general tools + state tools + state transitions
+     */
+    tools?: Array<
+      | State.EndCallTool
+      | State.TransferCallTool
+      | State.CheckAvailabilityCalTool
+      | State.BookAppointmentCalTool
+      | State.CustomTool
+    >;
+  }
+
+  export namespace State {
+    export interface Edge {
+      /**
+       * Describes what's the transition and at what time / criteria should this
+       * transition happen.
+       */
+      description: string;
+
+      /**
+       * The destination state name when going through transition of state via this edge.
+       * State transition internally is implemented as a tool call of LLM, and a tool
+       * call with name "transition*to*{destination_state_name}" will get created. Feel
+       * free to reference it inside the prompt.
+       */
+      destination_state_name: string;
+
+      /**
+       * After the state transitions, the agent would speak based on the new prompt and
+       * tools in the new state. This bit here controls whether to speak a transition
+       * sentence during the transition (so agent would say sentences like "Let's move on
+       * to the next section to help you set up an acount.", and state transitions, and
+       * agent continue to speak.). Usually this is not necessary, and is recommended to
+       * set to false to avoid LLM repeating itself during and after transition.
+       */
+      speak_during_transition: boolean;
+
+      /**
+       * Describes what parameters you want to extract out when the transition changes.
+       * The parameters extracted here can be referenced in prompts & function
+       * descriptions of later states via dynamic variables. 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.
+       */
+      parameters?: Edge.Parameters;
+    }
+
+    export namespace Edge {
+      /**
+       * Describes what parameters you want to extract out when the transition changes.
+       * The parameters extracted here can be referenced in prompts & function
+       * descriptions of later states via dynamic variables. 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.
+       */
+      export 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: Record<string, 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>;
+      }
+    }
+
+    export interface EndCallTool {
+      /**
+       * Name of the tool. Must be unique within all tools available to LLM at any given
+       * time (general tools + state tools + state transitions).
+       */
+      name: string;
+
+      type: 'end_call';
+
+      /**
+       * Describes when to end the call.
+       */
+      description?: string;
+    }
+
+    export interface TransferCallTool {
+      /**
+       * 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;
+
+      /**
+       * The number to transfer to in E.164 format (a + and country code, then the phone
+       * number with no space or other special characters). For example, +16175551212.
+       */
+      number: string;
+
+      type: 'transfer_call';
+
+      /**
+       * Describes when to transfer the call.
+       */
+      description?: string;
+    }
+
+    export interface CheckAvailabilityCalTool {
+      /**
+       * Cal.com Api key that have access to the cal.com event you want to check
+       * availability for.
+       */
+      cal_api_key: string;
+
+      /**
+       * Cal.com event type id number for the cal.com event you want to check
+       * availability for.
+       */
+      event_type_id: number;
+
+      /**
+       * Name of the tool. Must be unique within all tools available to LLM at any given
+       * time (general tools + state tools + state transitions).
+       */
+      name: string;
+
+      type: 'check_availability_cal';
+
+      /**
+       * Describes when to check availability.
+       */
+      description?: string;
+
+      /**
+       * Timezone to be used when checking availability, must be in
+       * [IANA timezone database](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones).
+       * If not specified, will check if user specified timezone in call, and if not,
+       * will use the timezone of the Retell servers.
+       */
+      timezone?: string;
+    }
+
+    export interface BookAppointmentCalTool {
+      /**
+       * Cal.com Api key that have access to the cal.com event you want to book
+       * appointment.
+       */
+      cal_api_key: string;
+
+      /**
+       * Cal.com event type id number for the cal.com event you want to book appointment.
+       */
+      event_type_id: number;
+
+      /**
+       * Name of the tool. Must be unique within all tools available to LLM at any given
+       * time (general tools + state tools + state transitions).
+       */
+      name: string;
+
+      type: 'book_appointment_cal';
+
+      /**
+       * Describes when to book the appointment.
+       */
+      description?: string;
+
+      /**
+       * Timezone to be used when booking appointment, must be in
+       * [IANA timezone database](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones).
+       * If not specified, will check if user specified timezone in call, and if not,
+       * will use the timezone of the Retell servers.
+       */
+      timezone?: string;
+    }
+
+    export interface CustomTool {
+      /**
+       * Describes what this tool does and when to call this tool.
+       */
+      description: string;
+
+      /**
+       * 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;
+
+      /**
+       * 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;
+
+      type: 'custom';
+
+      /**
+       * The URL we will post the function name and arguments to get a result for the
+       * function. Usually this is your server.
+       */
+      url: string;
+
+      /**
+       * 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;
+
+      /**
+       * 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;
+    }
+
+    export 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.
+       */
+      export 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: Record<string, 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>;
+      }
+    }
+  }
+}
+
+export type LlmListResponse = Array<LlmResponse>;
+
+export interface LlmCreateParams {
+  /**
+   * First utterance said by the agent in the call. If not set, LLM will dynamically
+   * generate a message. If set to "", agent will wait for user to speak first.
+   */
+  begin_message?: string;
+
+  /**
+   * General prompt that's appended to system prompt no matter what state the agent
+   * is in.
+   *
+   * - System prompt (with state) = general prompt + state prompt.
+   *
+   * - System prompt (no state) = general prompt.
+   */
+  general_prompt?: string;
+
+  /**
+   * A list of tools the model may call (to get external knowledge, call API, etc).
+   * You can select from some common predefined tools like end call, transfer call,
+   * etc; or you can create your own custom tool (last option) for the LLM to use.
+   *
+   * - Tools of LLM (with state) = general tools + state tools + state transitions
+   *
+   * - Tools of LLM (no state) = general tools
+   */
+  general_tools?: Array<
+    | LlmCreateParams.EndCallTool
+    | LlmCreateParams.TransferCallTool
+    | LlmCreateParams.CheckAvailabilityCalTool
+    | LlmCreateParams.BookAppointmentCalTool
+    | LlmCreateParams.CustomTool
+  >;
+
+  /**
+   * Name of the starting state. Required if states is not empty.
+   */
+  starting_state?: string;
+
+  /**
+   * States of the LLM. This is to help reduce prompt length and tool choices when
+   * the call can be broken into distinct states. With shorter prompts and less
+   * tools, the LLM can better focus and follow the rules, minimizing hallucination.
+   * If this field is not set, the agent would only have general prompt and general
+   * tools (essentially one state).
+   */
+  states?: Array<LlmCreateParams.State>;
+}
+
+export namespace LlmCreateParams {
+  export interface EndCallTool {
+    /**
+     * Name of the tool. Must be unique within all tools available to LLM at any given
+     * time (general tools + state tools + state transitions).
+     */
+    name: string;
+
+    type: 'end_call';
+
+    /**
+     * Describes when to end the call.
+     */
+    description?: string;
+  }
+
+  export interface TransferCallTool {
+    /**
+     * 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;
+
+    /**
+     * The number to transfer to in E.164 format (a + and country code, then the phone
+     * number with no space or other special characters). For example, +16175551212.
+     */
+    number: string;
+
+    type: 'transfer_call';
+
+    /**
+     * Describes when to transfer the call.
+     */
+    description?: string;
+  }
+
+  export interface CheckAvailabilityCalTool {
+    /**
+     * Cal.com Api key that have access to the cal.com event you want to check
+     * availability for.
+     */
+    cal_api_key: string;
+
+    /**
+     * Cal.com event type id number for the cal.com event you want to check
+     * availability for.
+     */
+    event_type_id: number;
+
+    /**
+     * Name of the tool. Must be unique within all tools available to LLM at any given
+     * time (general tools + state tools + state transitions).
+     */
+    name: string;
+
+    type: 'check_availability_cal';
+
+    /**
+     * Describes when to check availability.
+     */
+    description?: string;
+
+    /**
+     * Timezone to be used when checking availability, must be in
+     * [IANA timezone database](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones).
+     * If not specified, will check if user specified timezone in call, and if not,
+     * will use the timezone of the Retell servers.
+     */
+    timezone?: string;
+  }
+
+  export interface BookAppointmentCalTool {
+    /**
+     * Cal.com Api key that have access to the cal.com event you want to book
+     * appointment.
+     */
+    cal_api_key: string;
+
+    /**
+     * Cal.com event type id number for the cal.com event you want to book appointment.
+     */
+    event_type_id: number;
+
+    /**
+     * Name of the tool. Must be unique within all tools available to LLM at any given
+     * time (general tools + state tools + state transitions).
+     */
+    name: string;
+
+    type: 'book_appointment_cal';
+
+    /**
+     * Describes when to book the appointment.
+     */
+    description?: string;
+
+    /**
+     * Timezone to be used when booking appointment, must be in
+     * [IANA timezone database](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones).
+     * If not specified, will check if user specified timezone in call, and if not,
+     * will use the timezone of the Retell servers.
+     */
+    timezone?: string;
+  }
+
+  export interface CustomTool {
+    /**
+     * Describes what this tool does and when to call this tool.
+     */
+    description: string;
+
+    /**
+     * 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;
+
+    /**
+     * 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;
+
+    type: 'custom';
+
+    /**
+     * The URL we will post the function name and arguments to get a result for the
+     * function. Usually this is your server.
+     */
+    url: string;
+
+    /**
+     * 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;
+
+    /**
+     * 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;
+  }
+
+  export 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.
+     */
+    export 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: Record<string, 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>;
+    }
+  }
+
+  export interface State {
+    /**
+     * Name of the state, must be unique for each state.
+     */
+    name: string;
+
+    /**
+     * Edges of the state define how and what state can be reached from this state.
+     */
+    edges?: Array<State.Edge>;
+
+    /**
+     * Prompt of the state, will be appended to the system prompt of LLM.
+     *
+     * - System prompt = general prompt + state prompt.
+     */
+    state_prompt?: string;
+
+    /**
+     * A list of tools specific to this state the model may call (to get external
+     * knowledge, call API, etc). You can select from some common predefined tools like
+     * end call, transfer call, etc; or you can create your own custom tool (last
+     * option) for the LLM to use.
+     *
+     * - Tools of LLM = general tools + state tools + state transitions
+     */
+    tools?: Array<
+      | State.EndCallTool
+      | State.TransferCallTool
+      | State.CheckAvailabilityCalTool
+      | State.BookAppointmentCalTool
+      | State.CustomTool
+    >;
+  }
+
+  export namespace State {
+    export interface Edge {
+      /**
+       * Describes what's the transition and at what time / criteria should this
+       * transition happen.
+       */
+      description: string;
+
+      /**
+       * The destination state name when going through transition of state via this edge.
+       * State transition internally is implemented as a tool call of LLM, and a tool
+       * call with name "transition*to*{destination_state_name}" will get created. Feel
+       * free to reference it inside the prompt.
+       */
+      destination_state_name: string;
+
+      /**
+       * After the state transitions, the agent would speak based on the new prompt and
+       * tools in the new state. This bit here controls whether to speak a transition
+       * sentence during the transition (so agent would say sentences like "Let's move on
+       * to the next section to help you set up an acount.", and state transitions, and
+       * agent continue to speak.). Usually this is not necessary, and is recommended to
+       * set to false to avoid LLM repeating itself during and after transition.
+       */
+      speak_during_transition: boolean;
+
+      /**
+       * Describes what parameters you want to extract out when the transition changes.
+       * The parameters extracted here can be referenced in prompts & function
+       * descriptions of later states via dynamic variables. 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.
+       */
+      parameters?: Edge.Parameters;
+    }
+
+    export namespace Edge {
+      /**
+       * Describes what parameters you want to extract out when the transition changes.
+       * The parameters extracted here can be referenced in prompts & function
+       * descriptions of later states via dynamic variables. 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.
+       */
+      export 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: Record<string, 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>;
+      }
+    }
+
+    export interface EndCallTool {
+      /**
+       * Name of the tool. Must be unique within all tools available to LLM at any given
+       * time (general tools + state tools + state transitions).
+       */
+      name: string;
+
+      type: 'end_call';
+
+      /**
+       * Describes when to end the call.
+       */
+      description?: string;
+    }
+
+    export interface TransferCallTool {
+      /**
+       * 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;
+
+      /**
+       * The number to transfer to in E.164 format (a + and country code, then the phone
+       * number with no space or other special characters). For example, +16175551212.
+       */
+      number: string;
+
+      type: 'transfer_call';
+
+      /**
+       * Describes when to transfer the call.
+       */
+      description?: string;
+    }
+
+    export interface CheckAvailabilityCalTool {
+      /**
+       * Cal.com Api key that have access to the cal.com event you want to check
+       * availability for.
+       */
+      cal_api_key: string;
+
+      /**
+       * Cal.com event type id number for the cal.com event you want to check
+       * availability for.
+       */
+      event_type_id: number;
+
+      /**
+       * Name of the tool. Must be unique within all tools available to LLM at any given
+       * time (general tools + state tools + state transitions).
+       */
+      name: string;
+
+      type: 'check_availability_cal';
+
+      /**
+       * Describes when to check availability.
+       */
+      description?: string;
+
+      /**
+       * Timezone to be used when checking availability, must be in
+       * [IANA timezone database](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones).
+       * If not specified, will check if user specified timezone in call, and if not,
+       * will use the timezone of the Retell servers.
+       */
+      timezone?: string;
+    }
+
+    export interface BookAppointmentCalTool {
+      /**
+       * Cal.com Api key that have access to the cal.com event you want to book
+       * appointment.
+       */
+      cal_api_key: string;
+
+      /**
+       * Cal.com event type id number for the cal.com event you want to book appointment.
+       */
+      event_type_id: number;
+
+      /**
+       * Name of the tool. Must be unique within all tools available to LLM at any given
+       * time (general tools + state tools + state transitions).
+       */
+      name: string;
+
+      type: 'book_appointment_cal';
+
+      /**
+       * Describes when to book the appointment.
+       */
+      description?: string;
+
+      /**
+       * Timezone to be used when booking appointment, must be in
+       * [IANA timezone database](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones).
+       * If not specified, will check if user specified timezone in call, and if not,
+       * will use the timezone of the Retell servers.
+       */
+      timezone?: string;
+    }
+
+    export interface CustomTool {
+      /**
+       * Describes what this tool does and when to call this tool.
+       */
+      description: string;
+
+      /**
+       * 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;
+
+      /**
+       * 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;
+
+      type: 'custom';
+
+      /**
+       * The URL we will post the function name and arguments to get a result for the
+       * function. Usually this is your server.
+       */
+      url: string;
+
+      /**
+       * 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;
+
+      /**
+       * 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;
+    }
+
+    export 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.
+       */
+      export 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: Record<string, 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>;
+      }
+    }
+  }
+}
+
+export interface LlmUpdateParams {
+  /**
+   * First utterance said by the agent in the call. If not set, LLM will dynamically
+   * generate a message. If set to "", agent will wait for user to speak first.
+   */
+  begin_message?: string;
+
+  /**
+   * General prompt that's appended to system prompt no matter what state the agent
+   * is in.
+   *
+   * - System prompt (with state) = general prompt + state prompt.
+   *
+   * - System prompt (no state) = general prompt.
+   */
+  general_prompt?: string;
+
+  /**
+   * A list of tools the model may call (to get external knowledge, call API, etc).
+   * You can select from some common predefined tools like end call, transfer call,
+   * etc; or you can create your own custom tool (last option) for the LLM to use.
+   *
+   * - Tools of LLM (with state) = general tools + state tools + state transitions
+   *
+   * - Tools of LLM (no state) = general tools
+   */
+  general_tools?: Array<
+    | LlmUpdateParams.EndCallTool
+    | LlmUpdateParams.TransferCallTool
+    | LlmUpdateParams.CheckAvailabilityCalTool
+    | LlmUpdateParams.BookAppointmentCalTool
+    | LlmUpdateParams.CustomTool
+  >;
+
+  /**
+   * Name of the starting state. Required if states is not empty.
+   */
+  starting_state?: string;
+
+  /**
+   * States of the LLM. This is to help reduce prompt length and tool choices when
+   * the call can be broken into distinct states. With shorter prompts and less
+   * tools, the LLM can better focus and follow the rules, minimizing hallucination.
+   * If this field is not set, the agent would only have general prompt and general
+   * tools (essentially one state).
+   */
+  states?: Array<LlmUpdateParams.State>;
+}
+
+export namespace LlmUpdateParams {
+  export interface EndCallTool {
+    /**
+     * Name of the tool. Must be unique within all tools available to LLM at any given
+     * time (general tools + state tools + state transitions).
+     */
+    name: string;
+
+    type: 'end_call';
+
+    /**
+     * Describes when to end the call.
+     */
+    description?: string;
+  }
+
+  export interface TransferCallTool {
+    /**
+     * 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;
+
+    /**
+     * The number to transfer to in E.164 format (a + and country code, then the phone
+     * number with no space or other special characters). For example, +16175551212.
+     */
+    number: string;
+
+    type: 'transfer_call';
+
+    /**
+     * Describes when to transfer the call.
+     */
+    description?: string;
+  }
+
+  export interface CheckAvailabilityCalTool {
+    /**
+     * Cal.com Api key that have access to the cal.com event you want to check
+     * availability for.
+     */
+    cal_api_key: string;
+
+    /**
+     * Cal.com event type id number for the cal.com event you want to check
+     * availability for.
+     */
+    event_type_id: number;
+
+    /**
+     * Name of the tool. Must be unique within all tools available to LLM at any given
+     * time (general tools + state tools + state transitions).
+     */
+    name: string;
+
+    type: 'check_availability_cal';
+
+    /**
+     * Describes when to check availability.
+     */
+    description?: string;
+
+    /**
+     * Timezone to be used when checking availability, must be in
+     * [IANA timezone database](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones).
+     * If not specified, will check if user specified timezone in call, and if not,
+     * will use the timezone of the Retell servers.
+     */
+    timezone?: string;
+  }
+
+  export interface BookAppointmentCalTool {
+    /**
+     * Cal.com Api key that have access to the cal.com event you want to book
+     * appointment.
+     */
+    cal_api_key: string;
+
+    /**
+     * Cal.com event type id number for the cal.com event you want to book appointment.
+     */
+    event_type_id: number;
+
+    /**
+     * Name of the tool. Must be unique within all tools available to LLM at any given
+     * time (general tools + state tools + state transitions).
+     */
+    name: string;
+
+    type: 'book_appointment_cal';
+
+    /**
+     * Describes when to book the appointment.
+     */
+    description?: string;
+
+    /**
+     * Timezone to be used when booking appointment, must be in
+     * [IANA timezone database](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones).
+     * If not specified, will check if user specified timezone in call, and if not,
+     * will use the timezone of the Retell servers.
+     */
+    timezone?: string;
+  }
+
+  export interface CustomTool {
+    /**
+     * Describes what this tool does and when to call this tool.
+     */
+    description: string;
+
+    /**
+     * 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;
+
+    /**
+     * 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;
+
+    type: 'custom';
+
+    /**
+     * The URL we will post the function name and arguments to get a result for the
+     * function. Usually this is your server.
+     */
+    url: string;
+
+    /**
+     * 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;
+
+    /**
+     * 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;
+  }
+
+  export 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.
+     */
+    export 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: Record<string, 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>;
+    }
+  }
+
+  export interface State {
+    /**
+     * Name of the state, must be unique for each state.
+     */
+    name: string;
+
+    /**
+     * Edges of the state define how and what state can be reached from this state.
+     */
+    edges?: Array<State.Edge>;
+
+    /**
+     * Prompt of the state, will be appended to the system prompt of LLM.
+     *
+     * - System prompt = general prompt + state prompt.
+     */
+    state_prompt?: string;
+
+    /**
+     * A list of tools specific to this state the model may call (to get external
+     * knowledge, call API, etc). You can select from some common predefined tools like
+     * end call, transfer call, etc; or you can create your own custom tool (last
+     * option) for the LLM to use.
+     *
+     * - Tools of LLM = general tools + state tools + state transitions
+     */
+    tools?: Array<
+      | State.EndCallTool
+      | State.TransferCallTool
+      | State.CheckAvailabilityCalTool
+      | State.BookAppointmentCalTool
+      | State.CustomTool
+    >;
+  }
+
+  export namespace State {
+    export interface Edge {
+      /**
+       * Describes what's the transition and at what time / criteria should this
+       * transition happen.
+       */
+      description: string;
+
+      /**
+       * The destination state name when going through transition of state via this edge.
+       * State transition internally is implemented as a tool call of LLM, and a tool
+       * call with name "transition*to*{destination_state_name}" will get created. Feel
+       * free to reference it inside the prompt.
+       */
+      destination_state_name: string;
+
+      /**
+       * After the state transitions, the agent would speak based on the new prompt and
+       * tools in the new state. This bit here controls whether to speak a transition
+       * sentence during the transition (so agent would say sentences like "Let's move on
+       * to the next section to help you set up an acount.", and state transitions, and
+       * agent continue to speak.). Usually this is not necessary, and is recommended to
+       * set to false to avoid LLM repeating itself during and after transition.
+       */
+      speak_during_transition: boolean;
+
+      /**
+       * Describes what parameters you want to extract out when the transition changes.
+       * The parameters extracted here can be referenced in prompts & function
+       * descriptions of later states via dynamic variables. 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.
+       */
+      parameters?: Edge.Parameters;
+    }
+
+    export namespace Edge {
+      /**
+       * Describes what parameters you want to extract out when the transition changes.
+       * The parameters extracted here can be referenced in prompts & function
+       * descriptions of later states via dynamic variables. 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.
+       */
+      export 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: Record<string, 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>;
+      }
+    }
+
+    export interface EndCallTool {
+      /**
+       * Name of the tool. Must be unique within all tools available to LLM at any given
+       * time (general tools + state tools + state transitions).
+       */
+      name: string;
+
+      type: 'end_call';
+
+      /**
+       * Describes when to end the call.
+       */
+      description?: string;
+    }
+
+    export interface TransferCallTool {
+      /**
+       * 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;
+
+      /**
+       * The number to transfer to in E.164 format (a + and country code, then the phone
+       * number with no space or other special characters). For example, +16175551212.
+       */
+      number: string;
+
+      type: 'transfer_call';
+
+      /**
+       * Describes when to transfer the call.
+       */
+      description?: string;
+    }
+
+    export interface CheckAvailabilityCalTool {
+      /**
+       * Cal.com Api key that have access to the cal.com event you want to check
+       * availability for.
+       */
+      cal_api_key: string;
+
+      /**
+       * Cal.com event type id number for the cal.com event you want to check
+       * availability for.
+       */
+      event_type_id: number;
+
+      /**
+       * Name of the tool. Must be unique within all tools available to LLM at any given
+       * time (general tools + state tools + state transitions).
+       */
+      name: string;
+
+      type: 'check_availability_cal';
+
+      /**
+       * Describes when to check availability.
+       */
+      description?: string;
+
+      /**
+       * Timezone to be used when checking availability, must be in
+       * [IANA timezone database](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones).
+       * If not specified, will check if user specified timezone in call, and if not,
+       * will use the timezone of the Retell servers.
+       */
+      timezone?: string;
+    }
+
+    export interface BookAppointmentCalTool {
+      /**
+       * Cal.com Api key that have access to the cal.com event you want to book
+       * appointment.
+       */
+      cal_api_key: string;
+
+      /**
+       * Cal.com event type id number for the cal.com event you want to book appointment.
+       */
+      event_type_id: number;
+
+      /**
+       * Name of the tool. Must be unique within all tools available to LLM at any given
+       * time (general tools + state tools + state transitions).
+       */
+      name: string;
+
+      type: 'book_appointment_cal';
+
+      /**
+       * Describes when to book the appointment.
+       */
+      description?: string;
+
+      /**
+       * Timezone to be used when booking appointment, must be in
+       * [IANA timezone database](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones).
+       * If not specified, will check if user specified timezone in call, and if not,
+       * will use the timezone of the Retell servers.
+       */
+      timezone?: string;
+    }
+
+    export interface CustomTool {
+      /**
+       * Describes what this tool does and when to call this tool.
+       */
+      description: string;
+
+      /**
+       * 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;
+
+      /**
+       * 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;
+
+      type: 'custom';
+
+      /**
+       * The URL we will post the function name and arguments to get a result for the
+       * function. Usually this is your server.
+       */
+      url: string;
+
+      /**
+       * 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;
+
+      /**
+       * 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;
+    }
+
+    export 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.
+       */
+      export 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: Record<string, 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>;
+      }
+    }
+  }
+}
+
+export namespace Llm {
+  export import LlmResponse = LlmAPI.LlmResponse;
+  export import LlmListResponse = LlmAPI.LlmListResponse;
+  export import LlmCreateParams = LlmAPI.LlmCreateParams;
+  export import LlmUpdateParams = LlmAPI.LlmUpdateParams;
+}