Signa AI API
Add sign language to your product. Send text, a recording or live audio, and get a signing avatar back, as signs your app draws, a video, or a feed for your broadcast. The same engine as the Signa AI apps.
What you can build
Signing in your app or website
Text in, signs out a word at a time, so the avatar starts while the sentence is still arriving.
Signed videos
An MP4 of the avatar signing any text, in your colours, for social, e-learning or archives.
Live events
Stream a speaker's audio and get the signing back while they are still talking.
Subtitles
Timed subtitles from any recording or video, as SRT, VTT or JSON.
Television
The signer on air in vMix, CasparCG, OBS and more, from scripts or live studio sound.
Web pages and apps
Call Signa straight from a browser or app with short-lived client tokens.
How it works
- Words in. Text, or speech we transcribe first (Kinyarwanda, English, French, Swahili, Japanese).
- Signs out. Each word becomes a short clip of body and hand positions in Signa's compact
.sgnformat. Words that have been signed before come back almost instantly. - On screen. Play the signs in your own player, ask us for a video, or show the Signa avatar in your broadcast.
The basics
| Details | |
|---|---|
| Base URL | https://api.usesigna.org/v1 |
| Authentication | A bearer key: Authorization: Bearer sk_live_… |
| Requests | JSON, except audio uploads (multipart) and video exports (the .sgn file as the body) |
| Live | wss://api.usesigna.org/v1/live |
| Usage | Counted per month: characters signed, live minutes, videos and subtitle minutes |
Plans are agreed with you