Create a voice agent with model, voice, transcriber, and tool configuration.
/v1/agentsbearerAuth API token issued from the dashboard (Settings → API Tokens). Sent as Authorization: Bearer <token>.
In: header
application/json
TypeScript Definitions
Use the request body type in TypeScript.
application/json
application/json
application/json
application/json
Agent display name.
1 <= length <= 200Free-text agent description.
"draft" | "active" | "archived"LLM configuration.
Voice (text-to-speech) configuration.
Ambient background audio played under the agent’s voice.
Speech-to-text configuration.
Call-length and silence timeouts.
Voicemail detection and drop behavior for outbound calls.
Call transfer (warm/blind handoff) configuration. Supplying any of name/description/destinations/messageMode/customMessage/notifyOnTransfer replaces the whole stored config.
Retrieval-augmented generation (RAG) and knowledge-base access.
Cross-call contact memory configuration.
Calendar availability + booking configuration.
Post-call structured data extraction.
Outbound webhook, live event routing, and call-flow messaging.
Recording, logging, and compliance settings.
Turn-taking, interruption, and endpointing tuning.
Provider-specific tuning knobs. Each sub-object only applies when the matching provider is selected in model/voice/transcriber.
Tool ids to attach. Must belong to your organization (or be a shared built-in tool); replaces the full set.
items <= 50