hume 0.7.2 → 0.8.0

package/.mock/definition/empathic-voice/__package__.yml

@@ -55,7 +55,7 @@ types:
       version_type:
         type: string
         docs: >-
-          Indicates whether this tool is using a fixed version number or
+          Indicates whether this prompt is using a fixed version number or
           auto-updating to the latest version. Values from the VersionType enum.
       version_description:
         type: optional<string>
@@ -72,6 +72,64 @@ types:
       text:
         type: string
         docs: Text used for this version of the Prompt.
+  PostedCustomVoice:
+    docs: A custom voice specifications posted to the server
+    properties:
+      name:
+        type: string
+        docs: >-
+          String with the name of the voice to use. Maximum length of 75
+          characters. Will be converted to all-uppercase.
+      base_voice:
+        type: string
+        docs: The voice the custom voice is based off of.
+      speech_rate_multiplier:
+        type: optional<double>
+        docs: The speech rate multiplier for this custom voice.
+      parameter_model:
+        type: string
+        docs: >-
+          The name of the parameter model used to define which attributes are
+          used by `parameters`.
+      parameters:
+        type: optional<map<string, optional<double>>>
+        docs: Voice specification for a Config.
+  ReturnCustomVoice:
+    docs: A custom voice specification returned from the server
+    properties:
+      id:
+        type: string
+        docs: Identifier for a Custom Voice. Formatted as a UUID.
+      version:
+        type: integer
+        docs: >-
+          Version number for a Custom Voice. Version numbers should be integers.
+          The combination of custom_voice_id and version number is unique.
+      name:
+        type: string
+        docs: >-
+          String with the name of the voice to use. Maximum length of 75
+          characters. Will be converted to all-uppercase.
+      created_on:
+        type: integer
+        docs: The timestamp when the first version of this prompt was created.
+      modified_on:
+        type: integer
+        docs: The timestamp when this version of the prompt was created.
+      base_voice:
+        type: string
+        docs: The voice the custom voice is based off of.
+      speech_rate_multiplier:
+        type: optional<double>
+        docs: The speech rate multiplier for this custom voice.
+      parameter_model:
+        type: string
+        docs: >-
+          The name of the parameter model used to define which attributes are
+          used by `parameters`.
+      parameters:
+        docs: Voice specification for a Config.
+        type: map<string, double>
   PostedBuiltinTool:
     docs: A configuration of a built-in tool to be posted to the server
     properties:
@@ -101,6 +159,15 @@ types:
           Optional text that will be used as a verbatim event message. If the
           text is null and the event message is enabled, the system will
           generate a message based using the language model.
+  PostedEventMessageSpecs:
+    docs: Collection of event_message specs to be posted to the server
+    properties:
+      on_new_chat:
+        type: optional<PostedEventMessageSpec>
+        docs: >-
+          Set `enabled` to have the LLM generate an initial message using the
+          system prompt. Set `text` to a string to use verbatim as the initial
+          message.
   PostedLanguageModelModelProvider:
     enum:
       - OPEN_AI
@@ -108,6 +175,7 @@ types:
       - ANTHROPIC
       - FIREWORKS
       - GROQ
+      - GOOGLE
     docs: The provider of this model. Based on the enum modelProvider.
   PostedLanguageModel:
     docs: A LanguageModel to be posted to the server
@@ -141,6 +209,11 @@ types:
       duration_secs:
         type: optional<integer>
         docs: Duration in seconds for the timeout.
+  PostedTimeoutSpecs:
+    docs: Collection of timeout specs to be posted to the server
+    properties:
+      inactivity: optional<PostedTimeoutSpec>
+      max_duration: optional<PostedTimeoutSpec>
   PostedUserDefinedToolSpec:
     docs: A specific tool identifier to be posted to the server
     properties:
@@ -157,16 +230,20 @@ types:
       - ITO
       - DACHER
       - KORA
-    docs: String with the name of the voice to use.
+    docs: >-
+      String with the name of the voice to use. Maximum length of 75 characters.
+      Will be converted to all-uppercase.
   PostedVoice:
     docs: A Voice specification posted to the server
     properties:
       provider:
-        type: optional<literal<"HUME_AI">>
+        type: literal<"HUME_AI">
         docs: The provider of the voice to use. Based on the enum VoiceProvider.
       name:
-        type: PostedVoiceName
-        docs: String with the name of the voice to use.
+        type: optional<PostedVoiceName>
+        docs: >-
+          String with the name of the voice to use. Maximum length of 75
+          characters. Will be converted to all-uppercase.
   ReturnBuiltinTool:
     docs: A specific builtin tool version returned from the server
     properties:
@@ -212,12 +289,7 @@ types:
       builtin_tools:
         type: optional<list<optional<ReturnBuiltinTool>>>
         docs: List of built-in tools associated with this config
-      event_messages:
-        type: optional<map<string, optional<ReturnEventMessageSpec>>>
-        docs: Map of event messages associated with this config.
-      timeouts:
-        type: optional<map<string, optional<ReturnTimeoutSpec>>>
-        docs: Map of timeouts associated with this config.
+      event_messages: optional<ReturnEventMessageSpecs>
   ReturnEllmModel:
     docs: A specific eLLM Model configuration
     properties:
@@ -238,6 +310,15 @@ types:
           Optional text that will be used as a verbatim event message. If the
           text is null and the event message is enabled, the system will
           generate a message based using the language model.
+  ReturnEventMessageSpecs:
+    docs: Collection of event_message specs to be returned from the server
+    properties:
+      on_new_chat:
+        type: optional<ReturnEventMessageSpec>
+        docs: >-
+          `enabled` causes the LLM to generate an initial message using the
+          system prompt. `text` is the string to use verbatim as the initial
+          message.
   ReturnLanguageModel:
     docs: A specific LanguageModel
     properties:
@@ -259,15 +340,35 @@ types:
       duration_secs:
         type: optional<integer>
         docs: Duration in seconds for the timeout.
+  ReturnTimeoutSpecs:
+    docs: Collection of timeout specs to be returned from the server
+    properties:
+      inactivity: ReturnTimeoutSpec
+      max_duration: ReturnTimeoutSpec
   ReturnVoice:
     docs: A specific voice specification
     properties:
       provider:
         type: string
         docs: The provider of the voice to use. Based on the enum VoiceProvider.
+      name:
+        type: optional<string>
+        docs: >-
+          String with the name of the voice to use. Maximum length of 75
+          characters. Will be converted to all-uppercase.
+  ReturnPrivacySettings:
+    properties:
+      id: optional<string>
+      do_not_retain_data: optional<boolean>
+      do_not_use_for_training: optional<boolean>
+  PostedCustomVoiceName:
+    docs: A custom voice name change to be posted to the server
+    properties:
       name:
         type: string
-        docs: String with the name of the voice to use.
+        docs: >-
+          String with the name of the voice to use. Maximum length of 75
+          characters. Will be converted to all-uppercase.
   ReturnPagedUserDefinedTools:
     docs: A paginated list of user defined tool versions returned from the server
     properties:
@@ -298,6 +399,23 @@ types:
       prompts_page:
         docs: List of prompts returned for the specified page number and page size.
         type: list<optional<ReturnPrompt>>
+  ReturnPagedCustomVoices:
+    docs: A paginated list of custom voices returned from the server
+    properties:
+      page_number:
+        type: integer
+        docs: The page number of the returned results.
+      page_size:
+        type: integer
+        docs: The number of results returned per page.
+      total_pages:
+        type: integer
+        docs: The total number of pages in the collection
+      custom_voices_page:
+        docs: >-
+          List of custom voices returned for the specified page number and page
+          size.
+        type: list<ReturnCustomVoice>
   ReturnPagedConfigs:
     docs: A paginated list of config versions returned from the server
     properties:
@@ -609,92 +727,50 @@ types:
           List of chat_events returned for the specified page number and page
           size.
         type: list<ReturnChatEvent>
-  AssistantEnd:
-    docs: When provided, the output is an assistant end message.
+  ReturnPurgeStats:
     properties:
-      custom_session_id:
-        type: optional<string>
-        docs: >-
-          Used to manage conversational state, correlate frontend and backend
-          data, and persist conversations across EVI sessions.
-      type:
-        type: literal<"assistant_end">
-        docs: >-
-          The type of message sent through the socket; for an Assistant End
-          message, this must be `assistant_end`.
+      id: optional<string>
+      num_chats: optional<integer>
+      num_chat_events: optional<integer>
+      num_configs: optional<integer>
+      num_prompts: optional<integer>
+      num_tools: optional<integer>
+      num_voices: optional<integer>
   AssistantInput:
     docs: When provided, the input is spoken by EVI.
     properties:
-      custom_session_id:
-        type: optional<string>
-        docs: >-
-          Used to manage conversational state, correlate frontend and backend
-          data, and persist conversations across EVI sessions.
-      text:
-        type: string
-        docs: Text to be synthesized.
       type:
         type: literal<"assistant_input">
         docs: >-
           The type of message sent through the socket; for an Assistant Input
           message, this must be `assistant_input`.
-  AssistantMessage:
-    docs: When provided, the output is an assistant message.
-    properties:
       custom_session_id:
         type: optional<string>
         docs: >-
           Used to manage conversational state, correlate frontend and backend
           data, and persist conversations across EVI sessions.
-      from_text:
-        type: boolean
-        docs: Indicates if this message was constructed from a text input message.
-      id:
-        type: optional<string>
-        docs: ID of the assistant message.
-      message:
-        type: ChatMessage
-        docs: Transcript of the message.
-      models:
-        type: Inference
-        docs: Inference model results.
-      type:
-        type: literal<"assistant_message">
-        docs: >-
-          The type of message sent through the socket; for an Assistant Message,
-          this must be `assistant_message`.
+      text:
+        type: string
+        docs: Text to be synthesized.
   AudioConfiguration:
     properties:
-      channels:
-        type: integer
-        docs: Number of channels.
       encoding:
         type: Encoding
         docs: Audio encoding.
+      channels:
+        type: integer
+        docs: Number of channels.
       sample_rate:
         type: integer
         docs: Audio sample rate.
   AudioInput:
     docs: When provided, the input is audio.
     properties:
-      custom_session_id:
-        type: optional<string>
-        docs: >-
-          Used to manage conversational state, correlate frontend and backend
-          data, and persist conversations across EVI sessions.
-      data:
-        type: string
-        docs: Base64 encoded audio input.
-        validation:
-          format: base64
       type:
         type: literal<"audio_input">
         docs: >-
           The type of message sent through the socket; for an Audio Input
           message, this must be `audio_input`.
-  AudioOutput:
-    docs: When provided, the output is audio.
-    properties:
       custom_session_id:
         type: optional<string>
         docs: >-
@@ -702,236 +778,104 @@ types:
           data, and persist conversations across EVI sessions.
       data:
         type: string
-        docs: Base64 encoded audio output.
+        docs: Base64 encoded audio input.
         validation:
           format: base64
-      id:
-        type: string
-        docs: ID of the audio output.
-      type:
-        type: literal<"audio_output">
-        docs: >-
-          The type of message sent through the socket; for an Audio Output
-          message, this must be `audio_output`.
   BuiltInTool: literal<"web_search">
   BuiltinToolConfig:
     properties:
-      fallback_content: optional<string>
       name: BuiltInTool
-  ChatMessageToolResult:
-    discriminated: false
-    docs: Function call response from client.
-    union:
-      - ToolResponseMessage
-      - ToolErrorMessage
-  ChatMessage:
-    properties:
-      content:
-        type: optional<string>
-        docs: Transcript of the message.
-      role:
-        type: Role
-        docs: Role of who is providing the message.
-      tool_call:
-        type: optional<ToolCallMessage>
-        docs: Function call name and arguments.
-      tool_result:
-        type: optional<ChatMessageToolResult>
-        docs: Function call response from client.
-  ChatMetadata:
-    docs: When provided, the output is a chat metadata message.
-    properties:
-      chat_group_id:
-        type: string
-        docs: ID of the chat group. Used to resume a chat.
-      chat_id:
-        type: string
-        docs: ID of the chat.
-      custom_session_id:
-        type: optional<string>
-        docs: >-
-          Used to manage conversational state, correlate frontend and backend
-          data, and persist conversations across EVI sessions.
-      type:
-        type: literal<"chat_metadata">
-        docs: >-
-          The type of message sent through the socket; for a Chat Metadata
-          message, this must be `chat_metadata`.
+      fallback_content: optional<string>
   Context:
     properties:
-      text:
-        type: string
-        docs: User context to inject.
       type:
         type: optional<ContextType>
         docs: The persistence level of the injected context.
+      text:
+        type: string
+        docs: User context to inject.
   ContextType:
     enum:
       - editable
       - persistent
       - temporary
-  EmotionScores:
-    properties:
-      Admiration: double
-      Adoration: double
-      Aesthetic Appreciation: double
-      Amusement: double
-      Anger: double
-      Anxiety: double
-      Awe: double
-      Awkwardness: double
-      Boredom: double
-      Calmness: double
-      Concentration: double
-      Confusion: double
-      Contemplation: double
-      Contempt: double
-      Contentment: double
-      Craving: double
-      Desire: double
-      Determination: double
-      Disappointment: double
-      Disgust: double
-      Distress: double
-      Doubt: double
-      Ecstasy: double
-      Embarrassment: double
-      Empathic Pain: double
-      Entrancement: double
-      Envy: double
-      Excitement: double
-      Fear: double
-      Guilt: double
-      Horror: double
-      Interest: double
-      Joy: double
-      Love: double
-      Nostalgia: double
-      Pain: double
-      Pride: double
-      Realization: double
-      Relief: double
-      Romance: double
-      Sadness: double
-      Satisfaction: double
-      Shame: double
-      Surprise (negative): double
-      Surprise (positive): double
-      Sympathy: double
-      Tiredness: double
-      Triumph: double
   Encoding: literal<"linear16">
-  Error:
-    docs: When provided, the output is an error message.
+  ErrorLevel: literal<"warn">
+  PauseAssistantMessage:
+    docs: >-
+      Pause responses from EVI. Chat history is still saved and sent after
+      resuming. 
     properties:
-      code:
-        type: string
-        docs: Error code.
+      type:
+        type: literal<"pause_assistant_message">
+        docs: >-
+          The type of message sent through the socket; for a Pause Assistant
+          message, this must be `pause_assistant_message`.
       custom_session_id:
         type: optional<string>
         docs: >-
           Used to manage conversational state, correlate frontend and backend
           data, and persist conversations across EVI sessions.
-      message:
-        type: string
-        docs: Error message.
-      slug:
-        type: string
-        docs: Error slug.
+  ResumeAssistantMessage:
+    docs: >-
+      Resume responses from EVI. Chat history sent while paused will now be
+      sent. 
+    properties:
       type:
-        type: literal<"error">
+        type: literal<"resume_assistant_message">
         docs: >-
-          The type of message sent through the socket; for a Web Socket Error
-          message, this must be `error`.
-  ErrorLevel: literal<"warn">
-  Inference:
-    properties:
-      prosody:
-        type: optional<ProsodyInference>
-        docs: Prosody model inference results.
-  MillisecondInterval:
-    properties:
-      begin:
-        type: integer
-        docs: Start time of the interval in milliseconds.
-      end:
-        type: integer
-        docs: End time of the interval in milliseconds.
-  PauseAssistantMessage:
-    docs: >-
-      Pause responses from EVI. Chat history is still saved and sent after
-      resuming. 
-    properties:
+          The type of message sent through the socket; for a Resume Assistant
+          message, this must be `resume_assistant_message`.
       custom_session_id:
         type: optional<string>
         docs: >-
           Used to manage conversational state, correlate frontend and backend
           data, and persist conversations across EVI sessions.
+  SessionSettings:
+    docs: Settings for this chat session.
+    properties:
       type:
-        type: literal<"pause_assistant_message">
+        type: literal<"session_settings">
         docs: >-
-          The type of message sent through the socket; for a Pause Assistant
-          message, this must be `pause_assistant_message`.
-  ProsodyInference:
-    properties:
-      scores: EmotionScores
-  ResumeAssistantMessage:
-    docs: >-
-      Resume responses from EVI. Chat history sent while paused will now be
-      sent. 
-    properties:
+          The type of message sent through the socket; for a Session Settings
+          message, this must be `session_settings`.
       custom_session_id:
         type: optional<string>
         docs: >-
           Used to manage conversational state, correlate frontend and backend
           data, and persist conversations across EVI sessions.
-      type:
-        type: literal<"resume_assistant_message">
+      system_prompt:
+        type: optional<string>
         docs: >-
-          The type of message sent through the socket; for a Resume Assistant
-          message, this must be `resume_assistant_message`.
-  Role:
-    enum:
-      - assistant
-      - system
-      - user
-      - all
-      - tool
-  SessionSettings:
-    docs: Settings for this chat session.
-    properties:
-      audio:
-        type: optional<AudioConfiguration>
-        docs: Audio configuration.
-      builtin_tools:
-        type: optional<list<BuiltinToolConfig>>
-        docs: List of builtin tools to enable.
+          Instructions for how the system should respond to the user. Set to
+          null to use the default system prompt.
       context:
         type: optional<Context>
         docs: User context to inject. Set to null to disable context injection.
-      custom_session_id:
-        type: optional<string>
-        docs: >-
-          Used to manage conversational state, correlate frontend and backend
-          data, and persist conversations across EVI sessions.
+      audio:
+        type: optional<AudioConfiguration>
+        docs: Audio configuration.
       language_model_api_key:
         type: optional<string>
         docs: Third party API key for the language model used for non-Hume models.
-      system_prompt:
-        type: optional<string>
-        docs: >-
-          Instructions for how the system should respond to the user. Set to
-          null to use the default system prompt.
       tools:
         type: optional<list<Tool>>
         docs: List of tools to enable.
-      type:
-        type: literal<"session_settings">
-        docs: >-
-          The type of message sent through the socket; for a Session Settings
-          message, this must be `session_settings`.
+      builtin_tools:
+        type: optional<list<BuiltinToolConfig>>
+        docs: List of builtin tools to enable.
+      metadata: optional<map<string, unknown>>
   Tool:
     properties:
+      type:
+        type: ToolType
+        docs: Type of tool.
+      name:
+        type: string
+        docs: Name of the tool.
+      parameters:
+        type: string
+        docs: Parameters of the tool. Is a stringified JSON schema.
       description:
         type: optional<string>
         docs: Description of the function.
@@ -940,80 +884,45 @@ types:
         docs: >-
           Fallback content of the tool, passed to the LLM if the function call
           response fails.
-      name:
-        type: string
-        docs: Name of the tool.
-      parameters:
-        type: string
-        docs: Parameters of the tool. Is a stringified JSON schema.
-      type:
-        type: ToolType
-        docs: Type of tool.
-  ToolCallMessage:
-    docs: When provided, the output is a tool call.
+  ToolErrorMessage:
+    docs: When provided, the output is a function call error.
     properties:
+      type:
+        type: literal<"tool_error">
+        docs: >-
+          The type of message sent through the socket; for a Tool Error message,
+          this must be `tool_error`.
       custom_session_id:
         type: optional<string>
         docs: >-
           Used to manage conversational state, correlate frontend and backend
           data, and persist conversations across EVI sessions.
-      name:
-        type: string
-        docs: Name of the tool called.
-      parameters:
-        type: string
-        docs: Parameters of the tool call. Is a stringified JSON schema.
-      response_required:
-        type: boolean
-        docs: Whether a response is required from the developer.
-      tool_call_id:
-        type: string
-        docs: ID of the tool call.
       tool_type:
         type: optional<ToolType>
         docs: Type of tool called, either 'builtin' or 'function'.
-      type:
-        type: literal<"tool_call">
-        docs: >-
-          The type of message sent through the socket; for a Tool Call message,
-          this must be `tool_call`.
-  ToolErrorMessage:
-    docs: When provided, the output is a function call error.
-    properties:
-      code:
-        type: optional<string>
-        docs: Error code.
+      tool_call_id:
+        type: string
+        docs: ID of the tool call.
       content:
         type: optional<string>
         docs: The content passed to the LLM in place of the tool response.
-      custom_session_id:
-        type: optional<string>
-        docs: >-
-          Used to manage conversational state, correlate frontend and backend
-          data, and persist conversations across EVI sessions.
       error:
         type: string
         docs: Error message from the tool call, not exposed to the LLM or user.
+      code:
+        type: optional<string>
+        docs: Error code.
       level:
         type: optional<ErrorLevel>
         docs: Error level.
-      tool_call_id:
-        type: string
-        docs: ID of the tool call.
-      tool_type:
-        type: optional<ToolType>
-        docs: Type of tool called, either 'builtin' or 'function'.
-      type:
-        type: literal<"tool_error">
-        docs: >-
-          The type of message sent through the socket; for a Tool Error message,
-          this must be `tool_error`.
   ToolResponseMessage:
     docs: When provided, the output is a function call response.
     properties:
-      content:
-        type: string
-        docs: Return value of the tool call.
+      type:
+        type: literal<"tool_response">
+        docs: >-
+          The type of message sent through the socket; for a Tool Response
+          message, this must be `tool_response`.
       custom_session_id:
         type: optional<string>
         docs: >-
@@ -1022,13 +931,11 @@ types:
       tool_call_id:
         type: string
         docs: ID of the tool call.
+      content:
+        type: string
+        docs: Return value of the tool call.
       tool_name: optional<string>
       tool_type: optional<ToolType>
-      type:
-        type: literal<"tool_response">
-        docs: >-
-          The type of message sent through the socket; for a Tool Response
-          message, this must be `tool_response`.
   ToolType:
     enum:
       - builtin
@@ -1036,6 +943,11 @@ types:
   UserInput:
     docs: User text to insert into the conversation.
     properties:
+      type:
+        type: literal<"user_input">
+        docs: >-
+          The type of message sent through the socket; for a User Input message,
+          this must be `user_input`.
       custom_session_id:
         type: optional<string>
         docs: >-
@@ -1044,38 +956,259 @@ types:
       text:
         type: string
         docs: User text to insert into the conversation.
+  AssistantEnd:
+    docs: When provided, the output is an assistant end message.
+    properties:
       type:
-        type: literal<"user_input">
+        type: literal<"assistant_message">
         docs: >-
-          The type of message sent through the socket; for a User Input message,
-          this must be `user_input`.
-  UserInterruption:
-    docs: When provided, the output is an interruption.
+          The type of message sent through the socket; for an Assistant End
+          message, this must be `assistant_end`.
+      custom_session_id:
+        type: optional<string>
+        docs: >-
+          Used to manage conversational state, correlate frontend and backend
+          data, and persist conversations across EVI sessions.
+  AssistantMessage:
+    docs: When provided, the output is an assistant message.
     properties:
+      type:
+        type: literal<"assistant_message">
+        docs: >-
+          The type of message sent through the socket; for an Assistant Message,
+          this must be `assistant_message`.
       custom_session_id:
         type: optional<string>
         docs: >-
           Used to manage conversational state, correlate frontend and backend
           data, and persist conversations across EVI sessions.
-      time:
+      id:
+        type: optional<string>
+        docs: ID of the assistant message.
+      message:
+        type: ChatMessage
+        docs: Transcript of the message.
+      models:
+        type: Inference
+        docs: Inference model results.
+      from_text:
+        type: boolean
+        docs: Indicates if this message was constructed from a text input message.
+  AudioOutput:
+    docs: When provided, the output is audio.
+    properties:
+      type:
+        type: literal<"audio_output">
+        docs: >-
+          The type of message sent through the socket; for an Audio Output
+          message, this must be `audio_output`.
+      custom_session_id:
+        type: optional<string>
+        docs: >-
+          Used to manage conversational state, correlate frontend and backend
+          data, and persist conversations across EVI sessions.
+      id:
+        type: string
+        docs: ID of the audio output.
+      data:
+        type: string
+        docs: Base64 encoded audio output.
+        validation:
+          format: base64
+  ChatMessageToolResult:
+    discriminated: false
+    docs: Function call response from client.
+    union:
+      - ToolResponseMessage
+      - ToolErrorMessage
+  ChatMessage:
+    properties:
+      role:
+        type: Role
+        docs: Role of who is providing the message.
+      content:
+        type: optional<string>
+        docs: Transcript of the message.
+      tool_call:
+        type: optional<ToolCallMessage>
+        docs: Function call name and arguments.
+      tool_result:
+        type: optional<ChatMessageToolResult>
+        docs: Function call response from client.
+  ChatMetadata:
+    docs: When provided, the output is a chat metadata message.
+    properties:
+      type:
+        type: literal<"chat_metadata">
+        docs: >-
+          The type of message sent through the socket; for a Chat Metadata
+          message, this must be `chat_metadata`.
+      custom_session_id:
+        type: optional<string>
+        docs: >-
+          Used to manage conversational state, correlate frontend and backend
+          data, and persist conversations across EVI sessions.
+      chat_group_id:
+        type: string
+        docs: ID of the chat group. Used to resume a chat.
+      chat_id:
+        type: string
+        docs: ID of the chat.
+      request_id:
+        type: optional<string>
+        docs: ID of the initiating request.
+  EmotionScores:
+    properties:
+      Admiration: double
+      Adoration: double
+      Aesthetic Appreciation: double
+      Amusement: double
+      Anger: double
+      Anxiety: double
+      Awe: double
+      Awkwardness: double
+      Boredom: double
+      Calmness: double
+      Concentration: double
+      Confusion: double
+      Contemplation: double
+      Contempt: double
+      Contentment: double
+      Craving: double
+      Desire: double
+      Determination: double
+      Disappointment: double
+      Disgust: double
+      Distress: double
+      Doubt: double
+      Ecstasy: double
+      Embarrassment: double
+      Empathic Pain: double
+      Entrancement: double
+      Envy: double
+      Excitement: double
+      Fear: double
+      Guilt: double
+      Horror: double
+      Interest: double
+      Joy: double
+      Love: double
+      Nostalgia: double
+      Pain: double
+      Pride: double
+      Realization: double
+      Relief: double
+      Romance: double
+      Sadness: double
+      Satisfaction: double
+      Shame: double
+      Surprise (negative): double
+      Surprise (positive): double
+      Sympathy: double
+      Tiredness: double
+      Triumph: double
+  WebSocketError:
+    docs: When provided, the output is an error message.
+    properties:
+      type:
+        type: literal<"error">
+        docs: >-
+          The type of message sent through the socket; for a Web Socket Error
+          message, this must be `error`.
+      custom_session_id:
+        type: optional<string>
+        docs: >-
+          Used to manage conversational state, correlate frontend and backend
+          data, and persist conversations across EVI sessions.
+      code:
+        type: string
+        docs: Error code.
+      slug:
+        type: string
+        docs: Error slug.
+      message:
+        type: string
+        docs: Error message.
+  Inference:
+    properties:
+      prosody:
+        type: optional<ProsodyInference>
+        docs: Prosody model inference results.
+  MillisecondInterval:
+    properties:
+      begin:
         type: integer
-        docs: Unix timestamp of the detected user interruption.
+        docs: Start time of the interval in milliseconds.
+      end:
+        type: integer
+        docs: End time of the interval in milliseconds.
+  ProsodyInference:
+    properties:
+      scores: EmotionScores
+  Role:
+    enum:
+      - assistant
+      - system
+      - user
+      - all
+      - tool
+  ToolCallMessage:
+    docs: When provided, the output is a tool call.
+    properties:
+      name:
+        type: string
+        docs: Name of the tool called.
+      parameters:
+        type: string
+        docs: Parameters of the tool call. Is a stringified JSON schema.
+      tool_call_id:
+        type: string
+        docs: ID of the tool call.
+      type:
+        type: literal<"tool_call">
+        docs: >-
+          The type of message sent through the socket; for a Tool Call message,
+          this must be `tool_call`.
+      custom_session_id:
+        type: optional<string>
+        docs: >-
+          Used to manage conversational state, correlate frontend and backend
+          data, and persist conversations across EVI sessions.
+      tool_type:
+        type: optional<ToolType>
+        docs: Type of tool called, either 'builtin' or 'function'.
+      response_required:
+        type: boolean
+        docs: Whether a response is required from the developer.
+  UserInterruption:
+    docs: When provided, the output is an interruption.
+    properties:
       type:
         type: literal<"user_interruption">
         docs: >-
           The type of message sent through the socket; for a User Interruption
           message, this must be `user_interruption`.
+      custom_session_id:
+        type: optional<string>
+        docs: >-
+          Used to manage conversational state, correlate frontend and backend
+          data, and persist conversations across EVI sessions.
+      time:
+        type: integer
+        docs: Unix timestamp of the detected user interruption.
   UserMessage:
     docs: When provided, the output is a user message.
     properties:
+      type:
+        type: literal<"user_message">
+        docs: >-
+          The type of message sent through the socket; for a User message, this
+          must be `user_message`.
       custom_session_id:
         type: optional<string>
         docs: >-
           Used to manage conversational state, correlate frontend and backend
           data, and persist conversations across EVI sessions.
-      from_text:
-        type: boolean
-        docs: Indicates if this message was constructed from a text input message.
       message:
         type: ChatMessage
         docs: Transcript of the message.
@@ -1085,8 +1218,107 @@ types:
       time:
         type: MillisecondInterval
         docs: Start and End time of user message.
-      type:
-        type: literal<"user_message">
-        docs: >-
-          The type of message sent through the socket; for a User message, this
-          must be `user_message`.
+      from_text:
+        type: boolean
+        docs: Indicates if this message was constructed from a text input message.
+  TtsInput:
+    properties:
+      type: optional<literal<"tts">>
+  TextInput:
+    properties:
+      type: optional<literal<"text_input">>
+  FunctionCallResponseInput:
+    properties:
+      type: optional<literal<"function_call_response">>
+  ExtendedVoiceArgs:
+    properties:
+      text: string
+      voice_args: VoiceArgs
+  Features:
+    properties:
+      confidence:
+        type: optional<double>
+        docs: Confidence (F->M)
+        default: 0
+        validation:
+          min: -20
+          max: 20
+      enthusiasm:
+        type: optional<double>
+        docs: Enthusiasm (M->A)
+        default: 0
+        validation:
+          min: -20
+          max: 20
+      affirmative:
+        type: optional<double>
+        docs: Affirmative (F->A/M)
+        default: 0
+        validation:
+          min: -20
+          max: 20
+      articulate:
+        type: optional<double>
+        docs: Articulate (A->M)
+        default: 0
+        validation:
+          min: -20
+          max: 20
+      maturity:
+        type: optional<double>
+        docs: Maturity (M->F)
+        default: 0
+        validation:
+          min: -20
+          max: 20
+      defeated:
+        type: optional<double>
+        docs: Defeated (N->M)
+        default: 0
+        validation:
+          min: -20
+          max: 20
+      feminine:
+        type: optional<double>
+        docs: Feminine (M->F)
+        default: 0
+        validation:
+          min: -20
+          max: 20
+      nerdy:
+        type: optional<double>
+        docs: Nerdy (A->A)
+        default: 0
+        validation:
+          min: -20
+          max: 20
+  HTTPValidationError:
+    properties:
+      detail: optional<list<ValidationError>>
+  ValidationErrorLocItem:
+    discriminated: false
+    union:
+      - string
+      - integer
+  ValidationError:
+    properties:
+      loc: list<ValidationErrorLocItem>
+      msg: string
+      type: string
+  VoiceArgs:
+    properties:
+      voice: optional<VoiceNameEnum>
+      features: optional<Features>
+      characters_per_second_ratio:
+        type: optional<double>
+        default: 1
+        validation:
+          min: 0.5
+          max: 2
+      baseline: optional<boolean>
+      reconstruct: optional<boolean>
+  VoiceNameEnum:
+    enum:
+      - ITO
+      - KORA
+      - DACHER