Captions in your own language, made on your own device.
Norwegian and Swedish from the national libraries’ own models, and every other language Whisper knows. Nothing is uploaded. No account. No price per minute.
Google Play coming soon App Store coming soon
Whisper runs on the device. Nothing leaves it unless you share it.
Language first
NB-Whisper from the National Library of Norway for bokmål and nynorsk, KB-Whisper from the National Library of Sweden for Swedish. Trained on thousands of hours of their own language.
On your device
Speech recognition runs on your phone or PC (on the graphics card). No upload, no account, no price per minute.
A real editor behind it
Templates, fonts and animation for the captions. Multitrack editing, music, stills and slideshows.
The languages and the models
Each model is made by others and comes under its own licence. Thanks to all of them. Models marked “coming soon” are ready but not online yet; until then the multilingual model writes those languages.
| Language | Model | Made by | Licence |
|---|---|---|---|
| Norwegian (bokmål and nynorsk) | NB-Whisper | Nasjonalbiblioteket (NbAiLab) | Apache-2.0 |
| Swedish | KB-Whisper | KBLab, Kungliga biblioteket | Apache-2.0 |
| Danish coming soon | Røst-v3 | CoRal (Alexandra Instituttet, Alvenir) | CoRal model licence (OpenRAIL-M) |
| Icelandic coming soon | whisper-large-icelandic | Language and Voice Lab, Háskólinn í Reykjavík | CC BY 4.0 |
| Faroese coming soon | whisper-large-faroese | Carlos Mena, Ravnur / Language and Voice Lab | CC BY 4.0 |
| Finnish coming soon | whisper-large-finnish-v3 | Finnish-NLP | Apache-2.0 |
| Northern Sámi (a first draft) coming soon | Whisper Large Northern Sámi | Nasjonalbiblioteket (NbAiLab) | Apache-2.0 |
| Every other language | Whisper large-v3-turbo, medium, small | OpenAI (whisper.cpp: Georgi Gerganov) | MIT |