Skip to main content
A1 turns a starting image into a live character that listens, talks, and moves. Build a conversational companion, a presenter, or a tutor, then let users influence what happens next. Start with Avatar Playground. Run the template locally, upload your character’s image or choose an example, and go live. Use a Persona to shape how the character speaks, moves, and listens without building your own interface.

01 Get started

Use Avatar Playground to try the experience and find a setup you like. When you’re ready to build, the Quickstart shows how to create a session and connect its control channel. The player guides provide the video and audio connection.

02 Define the character

Choose the starting image and voice, then write a Persona for the character’s personality, speaking style, movement, and listening habits. A few sentences are enough to start.

03 Interaction and performance

Talk with the character, interrupt a reply, or give it exact words and actions to perform. Explore scripted performances, image changes for outfits and scenes, and natural object interaction.

04 Integrate your app

Manage session lifetimes and connect video and audio through TRTC or Agora. Configure speech recognition, the dialogue model, or speech synthesis as needed. If your application already generates reply audio, use pure rendering mode to drive the avatar with it.

05 API reference

Find REST request and response fields, plus the WebSocket events used to control a live session. Start with Create Session for the initial configuration, then use Client events and Server events for ongoing interaction.