Integration

Hume AI integration

Hume AI's voice APIs. Generate expressive speech from text with the Octave model (an audio file, or JSON with base64 audio, generation ids and word or phoneme timestamps), describe a new voice in words or pick one from the Voice Library, and save a generated voice for reuse. Manage the Empathic Voice Interface (EVI): configs, system prompts and function-calling tools, each versioned, and read past EVI chats with their transcripts, emotion scores and audio recordings (personal data of the people who spoke). Text to speech is metered per character against the plan. Not exposed: the live WebSocket chat and its control-plane messages, the streaming text-to-speech and voice conversion routes, and the Expression Measurement API (limited release on another host). Authentication is the account's API key from the Hume portal (API keys page), sent in the X-Hume-Api-Key header.

Connect Hume AI How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in Hume AI.Built from the 39 Hume AI tools below, on the connection you add once.

What you can do with Hume AI

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

Evi config version, 5 operations

create_evi_config_versionCreate evi config versionPOSTUpdates an EVI config by saving a new version (evi_version required); earlier versions stay. Which omitted settings carry over is not documented: read get_evi_config_version first and send what should stay.
delete_evi_config_versionDelete evi config versionDELETEDeletes one version of an EVI config; the other versions stay. This cannot be undone.
get_evi_config_versionGet evi config versionGETReturns one version of an EVI config with all its settings: voice, prompt, language model, tools, timeouts, event messages, turn detection and webhooks.
list_evi_config_versionsList evi config versionsGETLists the versions of one EVI config with their full settings. Pass the config id from list_evi_configs.
update_evi_config_version_descriptionUpdate evi config version descriptionPATCHChanges the description of one EVI config version; its settings do not change. Returns the version.

Evi prompt version, 5 operations

create_evi_prompt_versionCreate evi prompt versionPOSTUpdates an EVI system prompt by saving a new version with the text sent; earlier versions stay. Returns the new version.
delete_evi_prompt_versionDelete evi prompt versionDELETEDeletes one version of an EVI system prompt; the other versions stay. This cannot be undone.
get_evi_prompt_versionGet evi prompt versionGETReturns one version of an EVI system prompt with its text.
list_evi_prompt_versionsList evi prompt versionsGETLists the versions of one EVI system prompt with their text. Pass the prompt id from list_evi_prompts.
update_evi_prompt_version_descriptionUpdate evi prompt version descriptionPATCHChanges the description of one EVI system prompt version; its text does not change. Returns the version.

Evi tool version, 5 operations

create_evi_tool_versionCreate evi tool versionPOSTUpdates an EVI tool by saving a new version (parameters required); earlier versions stay. Returns the new version.
delete_evi_tool_versionDelete evi tool versionDELETEDeletes one version of an EVI tool; the other versions stay. This cannot be undone.
get_evi_tool_versionGet evi tool versionGETReturns one version of an EVI tool: name, description, JSON parameters and fallback content.
list_evi_tool_versionsList evi tool versionsGETLists the versions of one EVI tool with their definitions. Pass the tool id from list_evi_tools.
update_evi_tool_version_descriptionUpdate evi tool version descriptionPATCHChanges the description of one EVI tool version; its definition does not change. Returns the version.

Evi config, 4 operations

create_evi_configCreate evi configPOSTCreates an EVI config that voice chats can use: EVI version (3 or 4-mini), prompt, voice, language model, tools, timeouts, event messages and webhooks. Returns the config with its id.
delete_evi_configDelete evi configDELETEDeletes an EVI config and all its versions. Chats can no longer use it. This cannot be undone.
list_evi_configsList evi configsGETLists EVI (Empathic Voice Interface) configs, by default the latest version of each, with their voice, prompt, language model, tools, timeouts, event messages and webhooks. Filter by name.
rename_evi_configRename evi configPATCHRenames an EVI config; the name applies to all its versions. Returns an empty body.

Evi prompt, 4 operations

create_evi_promptCreate evi promptPOSTCreates an EVI system prompt (instructions that shape EVI's behavior, responses and style), to be used by a config. Returns the prompt with its id and version.
delete_evi_promptDelete evi promptDELETEDeletes an EVI system prompt and all its versions. This cannot be undone.
list_evi_promptsList evi promptsGETLists EVI system prompts, by default the latest version of each, with their text. Filter by name. Prompts shape how EVI behaves and speaks in chats that use a config pointing to them.
rename_evi_promptRename evi promptPATCHRenames an EVI system prompt; the name applies to all its versions. Returns an empty body.

Evi tool, 4 operations

create_evi_toolCreate evi toolPOSTCreates an EVI function-calling tool (name, description, JSON parameters) that a config can enable; your application runs it when EVI calls it. Returns the tool with its id.
delete_evi_toolDelete evi toolDELETEDeletes an EVI tool and all its versions. This cannot be undone.
list_evi_toolsList evi toolsGETLists EVI function-calling tools, by default the latest version of each, with their name, description and JSON parameters. Filter by name.
rename_evi_toolRename evi toolPATCHRenames an EVI tool; the name applies to all its versions. Returns an empty body.

Voice, 3 operations

create_voiceCreate voicePOSTSaves the voice of an earlier generation (generation_id from synthesize_speech) to your account under a name, for reuse in later speech requests. Returns the voice id.
delete_voiceDelete voiceDELETEDeletes one of your saved custom voices, by name. This cannot be undone.
list_voicesList voicesGETLists text-to-speech voices: provider HUME_AI returns Hume's shared Voice Library, CUSTOM_VOICE your saved voices. Pass a voice's name or id, with its provider, in synthesize_speech utterances.

Evi chat audio, 2 operations

get_evi_chat_audioGet evi chat audioGETReturns the audio reconstruction of a past EVI chat: status (QUEUED, IN_PROGRESS, COMPLETE, ERROR, CANCELLED) and, once COMPLETE, a signed download URL valid 60 minutes. The recording is personal data.
list_evi_chat_group_audioList evi chat group audioGETLists the audio reconstructions of the chats in one EVI chat group, each with its status and, once COMPLETE, a signed download URL valid 60 minutes. Recordings are personal data.

Evi chat event, 2 operations

list_evi_chat_eventsList evi chat eventsGETReturns one EVI chat with a page of its events: user and assistant messages (the transcript), tool calls and the emotion scores measured on speech. Personal data of the person who spoke.
list_evi_chat_group_eventsList evi chat group eventsGETReturns the events of all chats in one EVI chat group: user and assistant messages (the transcript), tool calls and emotion scores measured on speech. Personal data of the people who spoke.

Evi chat group, 2 operations

get_evi_chat_groupGet evi chat groupGETReturns one EVI chat group with a page of its chats: status, start and end times, event count and the config each used.
list_evi_chat_groupsList evi chat groupsGETLists EVI chat groups (a chat resumed from another joins its group) with chat count, first and most recent start times, most recent chat and config, filterable by config_id.

Speech, 2 operations

synthesize_speechSynthesize speechPOSTGenerates speech from text with Hume's Octave model and returns JSON: base64 audio per generation, its generation_id (for create_voice or context) and optional timestamps. Metered per character. Use synthesize_speech_file for a file.
synthesize_speech_fileSynthesize speech filePOSTGenerates speech from text with Hume's Octave model and returns the audio file (mp3 by default, or wav or pcm). Metered per character. Use synthesize_speech to get the generation_id or timestamps.

Evi chat, 1 operation

list_evi_chatsList evi chatsGETLists past EVI voice chats with status, start and end times, event count, the config used and the chat group, filterable by config and status. Use an id with list_evi_chat_events or get_evi_chat_audio.

Use Hume AI from Claude Code, Cursor or Codex

Connect Hume AI in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its Hume AI operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect Hume AI

Create your workspace
Sign up free. Nothing to install.
Connect Hume AI once
Add your Hume AI connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.