Bring conversations to life with Interactive Avatars
Build live, conversational avatars that engage visitors, help customers get answers faster, and guide buyers through your product in real time.
Live demos are subject to the Terms of Service.
Two ways to go live
Same avatars, same real-time technology. Launch quickly with Synthesia, or build on the stack you already use.
Configure it in Synthesia. Start helping customers on your website.
Set the name, instructions, voice and knowledge base, and go live in minutes. Synthesia runs the language model, speech, avatar and real-time media, so you can focus on the conversations that matter to your customers.
- Nothing to host, no keys in your page
- Visitors get answers without leaving the page
- Review conversations and track usage in your workspace
Your voice agent, your stack. A face your users can talk to.
Bring face-to-face interaction to your existing voice agent. Attach an AvatarSession to your LiveKit agent and keep your LLM, voice, tools, transport and UI. Synthesia streams a lip-synced avatar into your room.
- Keep the LLM, voice and tools that work for you
- Reach users on web and mobile, in your own UI
- LiveKit today, Pipecat coming soon
Let your coding agent handle integration
Get to your first conversation faster. Install the Synthesia skill and let your coding agent scaffold the LiveKit agent and configure your Interactive Avatar, with support for API keys and room setup.
Works with
Claude Code
Cursor
Codex
$ npx skills add synthesia-ai/skillsAdd an interactive avatar to our real-time support bot
What teams are building
Help customers get answers, buyers take the next step, and learners build confidence. Build with your own stack through the API, or let Synthesia run it all.
Inbound Sales & Demos
Help buyers explore your product and get the answers they need to take the next step.
Customer Support
Help customers resolve questions faster with always-on agents that answer from your docs and hand off to humans.
Training & Coaching
Help learners build confidence through real-time conversation and practice.
Kiosks & Concierge
Help visitors find answers and get where they need to go with a conversational guide.




Start in minutes, scale when you're ready
One credit balance, two rates. Pay only for conversation minutes. No commitment, no minimum, every Interactive Avatar model included.
Managed
Synthesia hosts the model, speech and avatar. Metered on workspace credits; sessions up to 45 minutes.
API
You bring the agent; we stream the avatar. Pay only for the minutes it's on screen.
Avatar videos vs. Interactive Avatars
Avatar videos
Built for scripted, one-way communication.
Turn a prepared script into a talking avatar video
Deliver the same message to every viewer
Play without listening or responding
Share and reuse the finished video anywhere
Interactive Avatars
Built for live, two-way conversation.
Create a real-time avatar that listens and responds
Build conversational experiences that adapt as users speak
Give your AI agent a human face and voice
Embed a live avatar in your product or website
Interactive Avatar FAQs
What can I create using Interactive Avatars?
Interactive Avatars can power conversational experiences across products, websites, internal tools, training flows, support journeys, kiosks, and any other use case where a real-time, human face makes the interaction better.
Can I bring my own LLM?
Yes. Interactive Avatars is designed for teams that want to connect their own LLM, agent, or conversational logic and use Synthesia to power the avatar layer. You keep full control of the conversation; we handle the face.
How does the integration work?
The Synthesia Interactive Avatar is a plugin component you attach to your LiveKit Agent. Your agent produces speech the way it always does. The component routes that audio to the hosted avatar worker, which renders an expressive Interactive Avatar and publishes it to your LiveKit room as a regular participant. No changes to your frontend are needed.
Is this a public API?
Yes. Interactive Avatars is now generally available. Anyone with a Synthesia API key can get started using the developer documentation.
How do I get access?
Create an API key in your Synthesia workspace, install the LiveKit plugin, and follow our quickstart to begin building.
What are the current limitations?
The plugin currently supports Python LiveKit Agents only. We're continuing to improve expressiveness, listening behaviors, interruption handling, and latency, and will share updates soon.

Ready to build with Interactive Avatars?
Connect your own agent or conversational logic to Synthesia's Interactive Avatars.