The private, local Swiss-army knife for voice & video on Windows — powered by local Whisper AI. Dictate into any app, transcribe YouTube / Skool / 1,800+ sites, and record meetings & your screen — all on your computer. No subscription. No account. No audio ever uploaded.
-brightgreen){kind=icon}
⬇ Download for Windows · Website · Report an issue
If EchoQuill saves you time, a ⭐ star helps more people find it.
- Dictation everywhere — hold a hotkey and talk; text lands in any app. Voice commands & rewrite-by-voice too.
- Video / URL transcription — YouTube (incl. Shorts), TikTok, Vimeo, Loom, Wistia and ~1,800 sites. Skool support built in: paste the lesson's video link or a signed
.m3u8(auto-adds the Referer), or sign in via your browser for member-only videos. - Keep the media — optional "Keep audio file" / "Keep video file" so you save the download, not just the transcript. Name any transcript yourself.
- Meeting / Record (Pro) — record what you HEAR on your PC (calls, webinars, any playing video), optionally your mic, and even capture the full screen as MP4 — then transcribe & summarize it locally. No link, no DevTools.
- AI, your way — cleanup, formatting & summaries via Ollama (100% local), Ollama Cloud, Claude, OpenAI, Groq, DeepSeek, Qwen or Z.AI. Off by default.
- Clips tray, editable history, favorites, search — click a clip to paste at your cursor; edit & re-save any past transcription; ★ favorite the good ones.
- Organized output — everything saves under
Documents/EchoQuill/(Transcriptions, Meetings, …). - Private & free — 100% local unless you enable a cloud provider. Keys stored in Windows Credential Manager. Free forever for dictation.
Speaking is about three times faster than typing — and the paid dictation apps know it, charging every month while routing your voice through their servers. EchoQuill does what they do, free and fully offline, with a few things they don't do at all.
| EchoQuill | Wispr Flow | Dragon | |
|---|---|---|---|
| Price | Free forever | $12+/mo | $200–700 |
| Works offline | Yes | No | Partly |
| Voice leaves your PC | Never | Yes | Configurable |
| Open source | Yes (MIT) | No | No |
| Live words as you speak | Yes | Yes | Yes |
| Voice-control your PC | Yes | No | Yes |
| Rewrite selected text by voice | Yes | Yes | Limited |
| Video / YouTube transcription | Yes | No | No |
| Batch-transcribe a list of URLs | Yes | No | No |
| Timestamped transcript search | Yes | No | No |
| Clipboard manager with drag & drop | Yes | No | No |
| Click a clip to paste at your cursor | Yes | No | No |
| Edit & manage your transcription history | Yes | No | No |
| Learning dictionary | Yes | Yes | Yes |
| Per-app tone & formatting | Yes | Yes | No |
| Choose your AI (Claude, OpenAI, Groq, Ollama…) | Yes | No | No |
| Light / dark / system theme | Yes | Yes | Yes |
| One-click self-update | Yes | Yes | Yes |
🎙 Live dictation, anywhere. Press Ctrl+Alt+Space, talk, and your words appear live in a floating pill, then land wherever your cursor is — email, Word, browsers, chat, code editors. Prefer push-to-talk? Hold Right Alt instead. Press Esc to cancel mid-sentence.
🎧 Voice-control your PC. Say "computer, open Chrome", "search for foreclosure listings", "press enter", "volume up", "lock the computer". Recognition is primed with the command vocabulary, and only a fixed, safe list of actions can ever run — a misheard word can't do anything destructive.
✍ Rewrite anything by voice. Highlight text in any app, press Ctrl+Alt+W, and say "make this more professional." The selection is rewritten in place.
🎬 Turn videos into text. Paste a link from YouTube (including Shorts), TikTok, and ~1,800 other sites, or pick any media file. Transcripts auto-save, named after the video, with title and URL on top. Batch mode chews through a whole list of URLs unattended, with a Stop button any time. Search a transcript and matches report when they were said: 4 matches — at 02:14, 05:37, 11:02.
📋 Clips & history you control. Recent dictations live in a draggable floating tray — click a clip and it pastes straight into whatever field your cursor is in, or drag it into a specific spot. A full transcription browser lets you search, multi-select, delete, and — unique among dictation apps — edit the text of any past transcription and save it back.
📖 A dictionary that learns. Names, jargon, and brands always come out right — add, edit, or remove entries, and repeated corrections become dictionary entries automatically.
🤖 Optional AI cleanup & formatting. Off by default. Plug in Anthropic Claude, OpenAI, Groq, Ollama (free, local), Ollama Cloud, DeepSeek, Qwen, or Z.AI (GLM) to structure your dictation — turn a spoken email into a proper email, or spoken notes into bullet points — with different behavior per app. Your keys are stored in Windows Credential Manager.
🎨 Your look. Dark, light, or follow-system theme.
🔒 Optional audio history. Keep a private local recording of each dictation, with a storage budget and one-click ZIP export. Off by default.
⬆ Updates itself. One click in Settings → About when a new version ships.
Plus: spoken punctuation, mic lock, media ducking, never-miss-the-first-word cueing, adjustable tail capture, daily/weekly/monthly stats, full data export, administrator-mode for elevated apps, and a first-run tour so you're dictating in under a minute.
- Download EchoQuill-Setup.exe from the latest release.
- Run it (Windows may warn about an unrecognized app — click More info → Run anyway; the entire source is right here for inspection).
- Press
Ctrl+Alt+Spaceand start talking.
Requirements: Windows 10/11 and a microphone. First dictation downloads a speech model (~140 MB) once; after that it's fully offline. An NVIDIA GPU makes it faster but isn't required.
- Voice, transcripts, history, dictionary: processed and stored only on your computer.
- Zero analytics, telemetry, or tracking — read the code, it's all here.
- Network is used only for one-time model downloads, video URLs you paste, update checks, and AI providers you enable with your own key.
Details: SECURITY.md
Free covers unlimited dictation forever, 5 video transcriptions, and your last 10 clips. Pro — unlimited video transcription, drag-and-drop file upload, an unlimited clip library with a Favorites tab, Ask AI about any video (answers grounded in the transcript, with timestamps), and priority support — is coming soon at $5/month or $39/year at echo-quill.com. Star + Watch this repo to catch the launch.
git clone https://github.com/CFRE-dotcom/echoquill.git
cd echoquill
install.bat # one-time dependency setup (Python 3.10+)
EchoQuill.bat # run from source
build_exe.bat # or build your own EchoQuill.exeReleases are compiled automatically by GitHub Actions from the tagged source — what you download is what you see here.
Issues and PRs welcome. One focused change per PR; open an issue first for anything non-trivial.
MIT — free for everyone, forever. Speech recognition: OpenAI Whisper via faster-whisper · Video downloads: yt-dlp · Inspired by FluidVoice for macOS (independent project, not affiliated).