EchoQuill is free, open-source voice dictation for Windows. Your words appear live as you talk — powered by local Whisper AI. No subscription, no account, and your voice never leaves your PC.
⬇ Download Free for Windows See Pro — $5/moBuilt as an open-source alternative to Wispr Flow and Dragon, with the features they charge for.
One hotkey. Words appear as you speak, then land wherever your cursor is — email, Word, browser, chat.
Local Whisper AI runs on your PC. No cloud, no account, no audio ever uploaded. Ever.
"Computer, open Chrome." "Volume up." "Select all." Control your PC hands-free with a safe, fixed command list.
Highlight any text, speak an instruction — "make this more professional" — and it's rewritten in place.
Paste a YouTube (or 1,800 other sites) link, or pick a file. Full transcript, auto-saved and searchable.
Your recent dictations in a floating tray. Click to paste at your cursor, or drag a clip into any window.
Names, jargon, brands — always right. It even learns from your repeated corrections automatically.
Plug in Claude, OpenAI, Groq, or free local Ollama for smarter cleanup and per-app tone. Off by default.
Words per day and time saved vs typing — stored only on your computer.
One or two emails a month. New features, new versions, and first dibs on Pro pricing.
Dictation is free forever. Pro removes the limits for power users.
Really free, and open source under MIT — you can read every line of code on GitHub. Pro exists for power users who want unlimited video transcription and a bigger clip library; that's the whole business model. No ads, no data selling.
No. Speech recognition runs entirely on your PC with OpenAI's open-source Whisper model. The only network use is the one-time model download — unless you deliberately enable a cloud AI provider with your own key.
Windows 10 or 11 and a microphone. Download, run, grant mic access, press Ctrl+Alt+Space, and talk. An NVIDIA GPU makes it faster but isn't required.
New open-source apps aren't code-signed (certificates cost hundreds per year). Click "More info → Run anyway." The code is public on GitHub for anyone to inspect — that's better transparency than a signature.
Whisper is the same family of models behind most modern transcription services. Pick your speed/accuracy tradeoff in Settings — from instant "tiny" to 99-language "maximum accuracy" for videos.