Talking Avatar
No avatar usage fees
There is no hosted avatar session to meter and no credit balance to refill. Run the free SDK for as long as you need, subject to your users' device resources.
MIT licensed · Free for commercial and noncommercial use
A free talking avatar.
Yours to build with.
Give your app a talking face with Talking Avatar Open Source. Turn your own 3D model into a talking avatar, driven by live audio or a recording, entirely in the browser. Free to use, modify and ship, for commercial and noncommercial projects.
No account. No API key. No credits. No subscription.
- SDK and model, forever
- $0
- Commercial and noncommercial use
- MIT
- Lip-sync model
- 1 MB
- Where the avatar runs
- Your browser
Free talking avatar software that runs in your browser
The SDK listens to audio, predicts mouth shapes and animates a 3D .glb avatar in real time. It plays the voice and moves the lips together. You bring the avatar and audio; the SDK handles lip sync, blinks and brows.
Any audio source
Use an MP3 or WAV file, your microphone, or a MediaStream from a WebRTC voice agent. The model follows audio without needing text or TTS-provider phoneme timings.
Open source, under your control
Read and modify the code, use the JavaScript API or React component, and self-host the lip-sync model. Lip-sync analysis processes audio locally in the browser.
How to create a talking avatar for free
- Try the 3D browser demo with a sample voice, your own recording or your microphone.
- Choose a compatible 3D .glb character. A talking mascot works if its model includes the required mouth shapes.
- Follow the installation guide to add the avatar to your website. Use your own recording or local TTS to avoid paid voice APIs.
The open-source talking avatar SDK guide includes npm installation, JavaScript and React examples, and self-hosting instructions.
Build a talking avatar into your website
If you want to add a talking avatar to a website without paying an avatar API provider, use the open-source SDK's JavaScript API. It runs on the client, so there is no hosted endpoint, API key or per-minute avatar charge.
Use live speech from your voice agent or play recorded audio. The open-source talking avatar API reference explains live streams, clip playback and how to connect your existing voice-agent integration.
Questions about the free talking avatar
Is this talking avatar completely free?
Yes. Talking Avatar Open Source has no subscription, credits, per-minute fees, watermarks or SDK usage quotas. The SDK and lip-sync model are MIT licensed. Your chosen voice service, avatar assets and website hosting may have their own costs and licences.
Can I use it in a commercial product?
Yes. The MIT License permits commercial and noncommercial use, modification and redistribution. Keep the copyright and permission notice with copies or substantial portions of the software, and check the licences of any avatar and audio assets you use.
Is there a free talking avatar API?
The free version provides a JavaScript API in your browser, rather than a hosted REST API. Use attachStream() for live audio and speak() for clips. You do not need an API key or an account. The separately hosted Realistic Avatars service is paid, at $0.02 a minute, and new accounts get $1 of free credits to try it.
Is this the same as Talking Avatar Realistic Avatars?
No. Talking Avatar Open Source is a free, MIT-licensed 3D talking avatar for the browser: it renders your own 3D .glb avatar and animates its face from any audio (lip sync from a 1 MB model, plus natural blinks and brows), with no server, account or API key. Talking Avatar Realistic Avatars is a hosted, paid service of ready-made photorealistic avatars, rendered on our servers and streamed to your page over WebRTC or into your LiveKit room, for $0.02 a minute; new accounts get $1 of free credits. They are separate products, and you don't need one to use the other.
Can I create a free text-to-speech talking avatar?
Yes. Generate speech with your chosen TTS engine and pass the resulting audio to speak() or attachStream(). The SDK animates the face; it does not generate speech or include an LLM. Use your own recordings or a local TTS engine if you want to avoid paid voice APIs.
Can I turn a photo into a talking video?
This SDK animates an existing 3D .glb avatar in a webpage. It does not turn a portrait photo into a video or export an MP4. Use the free demo to see the browser experience before integrating it.
Which avatars and browsers work?
Use a .glb avatar with the 15 Oculus viseme morph targets, or map your avatar's morph names with morphTargets. Current Chrome, Edge, Firefox and Safari support the required WebGL and Web Audio features. Playback needs a user interaction; live audio requires HTTPS or localhost.
Build with a free, open-source talking avatar
Try the real SDK in your browser, then use it in your own app. No signup or payment required.