Make any text or audio
speak in sign language.
Type, upload, or speak — Signa AI signs it back to you in real time. Built for classrooms, broadcasters, creators, and anyone who wants their content accessible to deaf audiences.
No signup required for the free tier. Kinyarwanda + English + 100+ spoken languages.
Four ways to feed the avatar
Pick the input that fits your moment. All four produce the same signed output.
Text
Paste a sentence and watch it sign back to you.
Audio file
MP3, WAV, M4A — transcribed and signed automatically.
Record
Tap the mic, speak, stop. We normalize, transcribe, sign.
Live
Speak continuously — every phrase signs the moment you pause.
An avatar that looks like your audience.
Pick skin tone, hair, clothing, gender, and age. The avatar renders cleanly at any size and is driven by the same MediaPipe-style landmark stream the signing community trusts.
- 6 skin tones × 6 hair colors × 6 shirts × 4 pants
- Male / female × child / teen / adult
- Choices persist across sessions
- Same renderer for embedded and standalone use
Who it's for
From a single Kinyarwanda lesson to a national broadcast feed.
Schools & teachers
Caption Kinyarwanda lessons for deaf students in real time. Free tier covers a classroom's daily needs.
Creators & YouTubers
Embed a signing avatar next to your video. Pro tier removes the watermark and unlocks higher monthly quota.
Broadcasters
National TV / live event captioning. Enterprise tier comes with SLAs, custom integration, and on-prem options.
One token per character. No surprises.
The free tier gets you started for nothing. Pro and Enterprise are character-bucketed monthly subscriptions paid via Paypack (MTN MoMo or Airtel Money) in RWF.
See full pricingFrequently asked
Does Signa support Kinyarwanda?
Yes — Kinyarwanda is our primary spoken language. We transcribe rw audio via Pindo's STT and Chrome's Web Speech API; both produce production-quality transcripts on real Kinyarwanda.
What about sign-language output for Kinyarwanda?
We route through sign.mt's pose model. Sign-language coverage varies by region, so for now we use sign.mt's American Sign Language renderer as the default while a Kinyarwanda-specific signed model is being built.
Can I embed Signa in my YouTube videos?
The Pro tier unlocks an embeddable widget you can drop next to a video. Enterprise unlocks custom integration including broadcast-grade keyed video output.
Is the avatar accurate to how a deaf signer would actually sign?
The pose data comes from sign.mt — a research-backed multi-language pose model. The avatar renders that pose data. Accuracy depends on the underlying model, which is improving rapidly.
What devices is Signa available on?
Web today (works in any modern browser; Chrome / Edge unlock the Live mode). Mobile, desktop, Chrome / Safari extensions are on the roadmap.
Sign your first sentence in under a minute.
No signup. No credit card. Just type and translate.
Open the translator