Open source • Local-first • Privacy-first

Don't type,
just speak.

The open-source voice-to-text app that turns your speech into clear, polished writing using local speech models.

macOS
Windows
Linux
OpenFlow app interface showing voice dictation in action
Keyboard

umm, hope your week has started well… I was talking to Cheyene earlier but reception was really bad and I think their going to handle the first part of the project, but I'm not totally sure. Also, I told the team the the new timeline should be ready by Friday, although it's probably going to slip.

Typed draft
OpenFlow

Hope your week is off to a good start. I was talking to Cheyene earlier, but the reception was really bad. I think they're going to handle the first part of the project, but I'm not totally sure. I also told the team the new timeline should be ready by Friday — although it might slip.

Voice draft
Local
On-device processing
Private
Zero telemetry
Open
GPL-3.0-only source
AI Auto-Edits

Speak naturally.
Land perfectly.

Rambled thoughts become clear, perfectly formatted text — without filler words, typos, or awkward pauses. OpenFlow transcribes and edits your voice locally, instantly.

Try it free
Raw Voice

umm so basically I think that, like, we should probably uh circle back on the project timeline

OpenFlow Output

I think we should circle back on the project timeline.

Personal Dictionary

Your words,
your way.

OpenFlow learns your unique vocabulary — names, jargon, acronyms — and spells them right every time. No more correcting the same word twice.

Add your words
OpenFlow Product
Kubernetes Tech
Nguyen Name
SaaS Acronym
PRD Acronym
Snippets

Voice shortcuts for
everything.

Create voice shortcuts for things you say over and over. Speak a cue like "calendar" and get your full scheduling link — no typing needed.

Create snippets
"calendar"
You can book a 30-minute call with me here: calendly.com/openflow
"hours"
Our support hours are Monday — Friday, 9 AM — 6 PM EST.
"intro"
Hi! I'm the founder of OpenFlow — the open-source, privacy-first voice dictation tool.
Multilingual

Speak in many
languages.

Choose a language supported by your selected local Whisper or Parakeet model. Language availability varies by backend and model.

Try your language
🇺🇸English
🇪🇸Español
🇫🇷Français
🇩🇪Deutsch
🇯🇵日本語
🇰🇷한국어
🇨🇳中文
🇵🇹Português
🇮🇳हिन्दी
🇸🇦العربية
🇷🇺Русский
🇹🇷Türkçe
Privacy First

Your voice never
leaves your machine.

OpenFlow processes recordings, transcripts, and cleanup locally with no telemetry. Optional model downloads, update checks, and project links use the internet only when requested.

Local by default

Speech processing runs on your device using local models. After setup, normal dictation works without an internet connection.

Local Whisper models

Choose from Tiny to Large models. Everything runs on your hardware, nothing in the cloud.

No telemetry or tracking

No analytics or behavioral tracking. Update checks and downloads never include your recordings, transcripts, or usage history.

Local AI cleanup

Optional local LLM for text cleanup. Run your own model server — or use rule-based cleanup — no data leaves your machine.

Full data ownership

Export, delete, or keep everything. Your dictionaries, snippets, and history are stored locally in SQLite.

Built for real work

Useful wherever you write

Keep recordings and transcripts on your machine while drafting in any application.

Privacy-first writing

Turn natural speech into structured emails, notes, and messages without a subscription.

Faster first drafts

Choose local Parakeet or Whisper backends that match your hardware and accuracy needs.

Local transcription

Reduce repetitive typing with configurable push-to-talk and hands-free dictation.

Accessible input

Expand spoken shortcuts into reusable templates for the text you write every day.

Reusable snippets

Audit the GPL-licensed source and verify how audio, text, storage, and networking behave.

Transparent by design

Keep recordings and transcripts on your machine while drafting in any application.

Privacy-first writing

Turn natural speech into structured emails, notes, and messages without a subscription.

Faster first drafts

Choose local Parakeet or Whisper backends that match your hardware and accuracy needs.

Local transcription

Reduce repetitive typing with configurable push-to-talk and hands-free dictation.

Accessible input

Expand spoken shortcuts into reusable templates for the text you write every day.

Reusable snippets

Audit the GPL-licensed source and verify how audio, text, storage, and networking behave.

Transparent by design

Open by design.
Ready for community.

OpenFlow is GPL-licensed source software. Audit the code, contribute improvements, or create an independent fork under the license terms. No vendor lock-in or subscription is required.

Open Source
Local Whisper
Zero Telemetry
Cross-Platform Builds

Start flowing.

Effortless local voice dictation, open source and private.
Your voice, your data, your choice of model.

Windows supported; macOS and Linux builds are experimental.