Retell AI
Retell AI is a platform to build, test, and deploy AI voice agents for phone calls. Its conversational voice API orchestrates LLM responses, telephony, speech, and call analysis so developers can place and receive natural-sounding automated calls, with a real-time WebSocket for bringing your own custom LLM and streaming call audio.
APIs
Agents
Create, retrieve, list, update, and delete voice agents that pair a response engine (Retell LLM, custom LLM, or conversation flow) with a voice and call-behavior settings.
Retell LLM
Manage the Retell-hosted LLM response engine - general prompt, model, temperature, tools, multi-state flows, begin message, and knowledge base references that drive agent conver...
Conversation Flow
Manage node/edge conversation flows for visually structured, deterministic agent conversations as an alternative response engine to the Retell LLM.
Phone Calls
Place outbound PSTN calls (POST /v2/create-phone-call) and retrieve or list completed calls with transcript, recording URL, call analysis, and cost breakdown.
Web Calls
Register browser-based (WebRTC) calls (POST /v2/create-web-call) and receive an access_token to hand to the Retell Web Client SDK for real-time in-browser voice conversations.
Phone Numbers
Buy Retell-managed numbers or import your own via SIP trunk, and bind inbound and outbound agents to each number for automated call handling.
Voices
List and retrieve the catalog of available voices across TTS providers (ElevenLabs, OpenAI, PlayHT, Cartesia, Deepgram) with accent, gender, age, and preview audio.
Knowledge Bases
Create knowledge bases from texts, files, and URLs, manage their sources, and reference them from a Retell LLM or conversation flow for retrieval-augmented answers during calls.
Batch Call
Queue outbound calling campaigns to many recipients from a single from_number with per-task dynamic variables, subject to available call concurrency.
Concurrency
Read the organization's current and maximum simultaneous-call concurrency (GET /get-concurrency) - the primary capacity and throttling lever for outbound and batch calling.
Collections
Retell AI API
OPENPricing Plans
Rate Limits
FinOps
Retell Finops
FINOPSEvent Specifications
Retell AI Custom LLM WebSocket
AsyncAPI 2.6 description of Retell AI's real-time **Custom LLM WebSocket** surface, documented at https://docs.retellai.com/api-references/llm-websocket. Unlike a one-way HTTP s...
ASYNCAPI