About EarScribe
EarScribe is a static web app that runs OpenAI's Whisper speech-recognition models directly in your browser using Transformers.js and the WebGPU / WebAssembly runtimes.
There is no backend that touches your audio. The Cloudflare Pages host only serves the static HTML, JS and CSS. The Whisper model weights are fetched from the Hugging Face CDN on first use and cached by your browser.
Built on Transformers.js, ONNX Runtime Web and the open-source Whisper models from OpenAI.
Free unlimited audio to text — FAQ
Is EarScribe really 100% free with no limits?
Yes — 100% free, unlimited, no sign-up and no login required. There are no paid tiers, no credit card, no per-minute fees, and no daily or monthly quota. The Whisper speech to text model runs on your own hardware, so there is no cloud compute cost to pass on. Free audio to text with no strings attached.
Is my audio private and secure?
Fully private. EarScribe is a local-first audio to text converter: your audio file never leaves your device and is never uploaded to any server. No account means no data collected about you. You can even use it offline once the model is downloaded.
Is my audio uploaded anywhere?
No. Free audio to text happens entirely locally — the audio is decoded on your device and fed to a Whisper model running in a Web Worker. Nothing is sent to a server.
Why is the first run slow?
The first time you pick a model for audio to text, your browser downloads it (80 MB to 800 MB depending on the size). After that the free audio to text model is cached and reused — no re-download needed.
What languages are supported?
Free audio to text supports 99 languages via Whisper. The language is detected automatically — you do not need to pick one.
Can I transcribe a video?
Not directly. For video audio to text, extract the audio first (any tool that can output MP3 or WAV will work) and drop that here.
Does this work offline?
After the model has been downloaded once, yes — free audio to text works offline. The page itself needs to be loaded first, but subsequent audio to text runs work without internet.
How do I convert an MP3 to text for free?
Drop the MP3 onto EarScribe for free MP3 to text. The file is decoded in your browser, Whisper transcribes it locally, and you can copy the MP3 to text output or export your MP3 to text as SRT, VTT, TXT or JSON. There is no upload step and no account required.
Is there a free Otter.ai or Rev alternative?
EarScribe is a free audio to text alternative to cloud transcription services like Otter.ai, Rev and HappyScribe. The trade-off: you do the compute on your own machine for free speech to text, but the audio never leaves it and there are no per-minute fees or subscriptions.
Can I generate SRT or VTT subtitles from audio?
Yes. After audio to text, click Export → .srt or .vtt. Cue timing comes from Whisper's emitted timestamps, so the subtitles align with the audio to text output without manual splitting.
Does this run OpenAI Whisper in the browser?
Yes — EarScribe runs free audio to text via Whisper (Tiny, Base, Small, and Large v3 Turbo) directly in the browser through Transformers.js on WebGPU, falling back to WebAssembly when unavailable.