Developers
Dictate commit messages, code comments, docs and long AI prompts without leaving your editor — VS Code, Cursor, the terminal, anywhere you type.
AI voice-to-text · 100% on-device · Free forever
Think out loud. We'll type it.
Private, on-device dictation in 29 languages. Speak in any app and your words appear instantly. No internet needed, and nothing leaves your machine. Mac today, Windows soon.
Free forever · No account · No internet needed
Model & performance
Built on Whisper, running fully on your device.
DijiFlow runs OpenAI's Whisper speech models entirely on your device — Apple's Neural Engine and GPU on Mac today, with AMD, Intel and NVIDIA GPUs coming for Windows. 29 languages, fully offline. Nothing is ever uploaded.
Works offline · no internet neededmacOS 14 (Sonoma)+ · Apple SiliconWindows soon · AMD · Intel · NVIDIA GPUs29 languages, fully on-device
~98%accuracy on clear speech (largest model)
15×faster than real time on Apple Silicon
29languages, offline
No servers100% on your device
300 MB–6 GBspeech model — base to Large-v3 (downloaded once)
~5 MBapp size — speech models download on demand
Who it's for
However you work, talk it into text.
From code to meetings, DijiFlow turns speech into text on your own hardware. Fast, private, free to start.
Podcasters & creators
Transcribe episodes and interviews locally, with speaker labels and timestamps on Pro. No uploads and no per-minute cloud fees.
AI & LLM workflows
Speak your prompts instead of typing them. Long, accurate, punctuated text straight into ChatGPT, Claude and Cursor — nothing logged anywhere.
Meetings & research
Record and transcribe meetings, lectures and interviews privately. Speaker identification tracks who said what — all on your machine.
Writers & students
Draft articles, essays, notes and emails at the speed of speech in any app, with vocabulary packs for the names and jargon of your field.
Email & messaging
Reply by voice in Gmail, Outlook, WhatsApp, Telegram, Slack and Messenger. One hotkey dictates into whatever window has focus, so you never switch tools.
Teams & business
Roll out private dictation across your org with volume and company licences. No audio leaves your people's machines — built for regulated industries. Business licensing →
How it works
Three steps, zero friction.
Press a hotkey, speak, and your words land as text wherever your cursor is — fully on-device.
1
Press the hotkey
Hit Right Option + Space from anywhere on your computer.
2
Speak naturally
Talk the way you think — in any app, in any of 29 languages.
3
Your text appears
Pasted at your cursor instantly — nothing ever leaves the device.
Productivity
Talk is faster than typing.
Most people speak three to four times faster than they type. DijiFlow turns that into finished text, and none of it leaves your computer.
≈3.7×
faster than typing — ~150 wpm spoken vs ~40 wpm typed
~3 hrs
saved a week — at 2,000 words/day, speaking takes ~13 min vs ~50 min typed
~93%
less time on a 1-hour meeting transcript — ~4 min, not an hour
£0
to start — Free forever, no account, no card
Pricing
Start free. Go Pro when you're ready.
Free forever, no account. Upgrade to Pro for unlimited length, the largest models and speaker labels — one app, pick your term.
Free
£0
forever
No account, no card. Yours to keep.
Volume licensing · central billing & one invoice · priority support · buy once, use forever
Hotkey slots1 slot3 slots
File length3 minUnlimited
File & video transcription—Yes
Speaker labels & timestamps—Yes
Vocabulary packs—Yes
GPU / Neural Engine + largest models—Yes
All 29 languagesYesYes
Private by design
Your words never leave your machine.
Every word is transcribed on your own device. Nothing is uploaded, nothing sits on a server, and you don't need an internet connection once you're set up.
- No internet needed. It works completely offline
- Your audio and transcripts stay on your machine
- Keep a local history on your computer, only if you want one
- No account, no telemetry, no cloud copy
Questions
Frequently asked questions
Yes — the Free tier is free forever, with no account or card. Pro adds more slots, unlimited file/video, speaker labels and the largest models, with a 30-day free trial.
No. Transcription runs 100% on your machine. The internet is only used once, to download a speech model the first time.
Windows is coming soon. DijiFlow is on Mac today (macOS 14+, Apple Silicon); your licence will cover Windows when it launches.
Only to detect your global hotkey (default Right Option + Space). DijiFlow does not read what you type in other apps.
Talk instead of type
Free forever to start.
Download DijiFlow Dictate free for Mac, or try every Pro feature free for 30 days. Windows coming soon.
Free forever · No account · No internet needed