Trillet AI logo

Agents and voices

Create and configure voice agents and pick voices.

Create and configure voice agents and pick voices. 6 tools.

trillet_list_agents

List agents

List all voice agents in the workspace with a compact summary (id, name, LLM, voice, linked call flow).

Flags: read-only structured output

Parameters

NameTypeRequiredDescription
workspace_idstringnoRun this call in another workspace: its id or its name (see trillet_list_workspaces). Omit to use the default (TRILLET_WORKSPACE_ID, the workspace picked when the connector was approved, or the one set with trillet_switch_workspace).

Structured output

Returned as structuredContent (fields are optional; the same data is in the text result, where it is a bare array):

FieldTypeDescription
agentsobject[]

trillet_get_agent

Get agent

Fetch one agent's full configuration by id.

Flags: read-only structured output

Parameters

NameTypeRequiredDescription
agent_idstringyes
workspace_idstringnoRun this call in another workspace: its id or its name (see trillet_list_workspaces). Omit to use the default (TRILLET_WORKSPACE_ID, the workspace picked when the connector was approved, or the one set with trillet_switch_workspace).

Structured output

Returned as structuredContent (fields are optional; the same data is in the text result):

FieldTypeDescription
_idstring
namestring
llmModelstring
ttsModelany
pathwayany

trillet_create_agent

Create agent

Create a new voice agent. ttsModel is REQUIRED by the API (requests without it fail with a 500). Creating an agent usually also creates a linked call flow; check the response's pathway field. Sensible defaults: Rime voice, llmModel "platform-llm".

Flags: writes

Parameters

NameTypeRequiredDescription
namestringyesAgent display name
llmModelstringnoLLM id, e.g. "platform-llm", "gemini-2.5-flash", "gpt-5.2" Default: "platform-llm".
ttsModelobjectyesText-to-speech voice configuration
workspace_idstringnoRun this call in another workspace: its id or its name (see trillet_list_workspaces). Omit to use the default (TRILLET_WORKSPACE_ID, the workspace picked when the connector was approved, or the one set with trillet_switch_workspace).

trillet_update_agent

Update agent

Update an agent's name, LLM model, voice, settings, or TTS pronunciation dictionary. Only send the fields you want to change. ttsDictionary is the agent-level pronunciation list (ElevenLabs voices): the API syncs it to the voice provider and manages the dictionary locator ids itself; send only words.

Flags: writes idempotent

Parameters

NameTypeRequiredDescription
agent_idstringyes
namestringno
llmModelstringno
ttsModelobjectnoText-to-speech voice configuration
ttsDictionaryobjectnoAgent-level TTS pronunciation dictionary (agent.ttsDictionary). Applies to ElevenLabs voices; other providers keep the words but do not sync them.
settingsobjectno
workspace_idstringnoRun this call in another workspace: its id or its name (see trillet_list_workspaces). Omit to use the default (TRILLET_WORKSPACE_ID, the workspace picked when the connector was approved, or the one set with trillet_switch_workspace).

trillet_delete_agent

Delete agent

Delete an agent from the workspace. Unbind its phone numbers first (a bound number stops routing). Always confirm with the user; requires confirm=true. The linked call flow is a separate document and is not deleted.

Flags: writes destructive asks for confirmation

Clients that support MCP elicitation show a confirmation form before this tool runs; declining cancels it with no changes. Other clients rely on their own approval prompt and the confirm argument.

Parameters

NameTypeRequiredDescription
agent_idstringyes
confirmbooleanyesMust be true. Only set after the user explicitly approved deleting this agent.
workspace_idstringnoRun this call in another workspace: its id or its name (see trillet_list_workspaces). Omit to use the default (TRILLET_WORKSPACE_ID, the workspace picked when the connector was approved, or the one set with trillet_switch_workspace).

trillet_search_voices

Search voices

Search the built-in voice catalog across providers (Rime, ElevenLabs, Cartesia, OpenAI, Hume). All filters optional; with no filters returns the full catalog (trimmed to limit). Use the returned provider + voiceId in an agent's ttsModel.

Flags: read-only structured output

Parameters

NameTypeRequiredDescription
namestringno
providerstringno
genderstringno
accentstringnoe.g. "australian", "british", "american"
categorystringno
agestringno
use_casestringno
limitnumbernoMax results returned (applied client-side) Default: 50.
workspace_idstringnoRun this call in another workspace: its id or its name (see trillet_list_workspaces). Omit to use the default (TRILLET_WORKSPACE_ID, the workspace picked when the connector was approved, or the one set with trillet_switch_workspace).

Structured output

Returned as structuredContent (fields are optional; the same data is in the text result):

FieldTypeDescription
totalnumber
returnednumber
voicesobject[]