Live APIs (Beta)

Enable real-time, low-latency interaction with Convai AI characters through WebRTC-based Live APIs that support text, audio, and dynamic contextual communication.

Live APIs are available only on the Enterprise Plan.

Convai’s Live APIs enable real-time, dynamic interaction between users and AI characters through WebRTC. These APIs allow your applications to send and receive audio, text, and other contextual data, making conversations with AI characters seamless and responsive.

Once your characters are created, the Live APIs manage conversational input and contextual responses automatically, ensuring natural and consistent behavior across sessions.

With the Live APIs, you can:

• Engage in live conversations — Send user text or voice inputs and receive contextual, natural replies in real time.
• Retrieve responses in text and/or audio form — Power both text-based chat and voice dialogue experiences.
• Use realtime models — Integrate with low-latency models (currently Google Realtime; OpenAI support coming soon) for high-speed conversational performance.

The Live APIs act as the bridge between your AI characters and users, enabling smooth, two-way communication across both text and voice channels — all in real time.

---

Where to start

Page	What it covers
Connect API	Opening a session and every option you can configure on it
Use multi-character sessions	Creating a shared room with multiple character instances, routing turns, and updating the roster
Build a group chat	Running a moderated room where several characters answer the same message, in one connection
Group chat messages	The group-address command, its responses and error codes, and the turn-complete event
Turn lifecycle and message ordering	How a bot turn is delivered, which ordering you can rely on, and field presence rules
Response contract and parsing	How speech, actions, and emotion are separated — and exactly what the server removes from the spoken response
Message Glossary	Every message type at a glance
Client-to-server messages	Messages you send
Server-to-client messages	Messages you receive
Audio Data (via data channel)	Custom audio handling
Metrics	Performance metrics and monitoring

If you are writing a client from scratch, read Turn lifecycle and message ordering first. It explains the three envelope shapes on the data channel and the ordering guarantees, which together account for most integration issues.