Blog ShipAny
Blog
Przeczytaj najnowsze funkcje produktów, rozwiązania i aktualizacje.
Dolphin Is Now Available on EveryScribe — 40 Asian Languages and 22 Chinese Dialects in 238 MB
Dolphin, the compact multilingual ASR model covering 40 Eastern languages and 22 Chinese dialects, is now available on EveryScribe's Private Transcriber. Run it entirely in your browser at just 238 MB.
FunASR Nano Is Now Available on EveryScribe — LLM-Powered ASR for Chinese, Japanese, and Dialects
FunASR Nano — an LLM-enhanced ASR model powered by Qwen3-0.6B — is now available on EveryScribe's Private Transcriber. With support for 7 Chinese dialects, 26 regional accents, English, and Japanese, this 980 MB model delivers the highest accuracy in our lineup.
Giga-AM v2 Is Now Available on EveryScribe — The Best Open-Source Russian ASR Model
SberDevices' Giga-AM v2, the leading open-source Russian speech recognition model trained on 50,000 hours of Russian audio, is now available on EveryScribe's Private Transcriber. 225 MB, fully private, runs in your browser.
Introducing EveryScribe: Unlimited Transcription, Built for Everyone
Today we're officially launching EveryScribe — a privacy-first transcription platform that runs entirely in your browser. Here's why we built it, and where we're taking it.
Moonshine Base Is Now Available on EveryScribe — The Fastest English ASR Model for Your Browser
Moonshine Base by Useful Sensors is now available on EveryScribe's Private Transcriber. At just 275 MB, this edge-optimized English ASR model is up to 5× faster than Whisper and runs entirely in your browser with zero data uploads.
Omnilingual 1B Is Now Available on EveryScribe — ASR for 1,600+ Languages in Your Browser
Meta's Omnilingual 1B model — supporting over 1,600 languages including 500+ previously unsupported by AI — is now available on EveryScribe's Private Transcriber. Run it entirely in your browser with full privacy.
Omnilingual 300M Is Now Available on EveryScribe — 1,600 Languages at 348 MB
Meta's Omnilingual 300M model brings 1,600+ language speech recognition to EveryScribe's Private Transcriber in a lighter 348 MB package. Fast, private, and runs entirely in your browser.
Parakeet TDT-CTC 0.6B for Japanese Is Now Available on EveryScribe — NVIDIA's Dedicated Japanese ASR Model
NVIDIA's Parakeet TDT-CTC 0.6B Japanese model is now available on EveryScribe's Private Transcriber. With 6.4% CER on JSUT and full punctuation support, this 625 MB model delivers state-of-the-art Japanese transcription in your browser.
Parakeet TDT 0.6B Is Now Available on EveryScribe — NVIDIA's High-Accuracy ASR for English and 25 European Languages
NVIDIA's Parakeet TDT 0.6B v3 is now available on EveryScribe's Private Transcriber. With 1.69% WER on LibriSpeech and support for 25 European languages, this 640 MB model runs privately in your browser.
SenseVoice Is Now Available on EveryScribe — Fast, Multilingual ASR for East Asian Languages
SenseVoice by Alibaba FunAudioLLM is now available on EveryScribe's Private Transcriber. Learn why this 250 MB model delivers 7–15× faster transcription than Whisper for Chinese, Japanese, Korean, Cantonese, and English — running entirely in your browser.