ok so um can you tell the team the launch is is slipping to thursday, we're still waiting on legal uh, and i'll send the new timeline tonight
Can you tell the team the launch is slipping to Thursday? We're still waiting on legal, and I'll send the new timeline tonight.
This is the whole app.
Hold. Talk. Let go.
Every text field, every app.
Mail · Slack · Notes · Cursor · Claude Code · your browser
If a cursor blinks there, Sayly types there.
Your Mac does the work.
Speech and cleanup models run on your Mac. Nothing leaves it unless you say so.
Said once. Written for the room.
ok so um can you tell the team the launch is is slipping to thursday? we're still waiting on legal, uh, and i'll send the new timeline tonight
Why it's free
I built Sayly because every dictation app charges for things your Mac already does.
Sayly started as a tool for me. I wanted to hold a key, say a sentence, and have clean text land where my cursor was, without paying a subscription for work my Mac can do on its own. So the speech model and the cleanup model run locally, and every feature is free: models, your own API keys, modes, vocabulary, history, updates. The one thing I charge for is Sayly Cloud, a hosted API for people who would rather not manage keys or models. If you never need it, you never pay.
Burak Ünlü, maker of Sayly
Free. Actually.
Free
$0
Every feature. No account, no trial, no card.
- Local speech and cleanup models
- Your own API keys
- Your own GGUF models
- Modes
- Vocabulary
- History
- Updates
- No telemetry
- Audio kept 14 days by default, then deleted; you can turn storage off
Sayly Cloud
$9/mo
Hosted speech and cleanup, nothing to set up.
- Hosted speech and cleanup
- No setup
- Fastest
- Fair-use cap
Questions
Is it really free?
Yes. Every feature is free, for good: the local models, your own API keys, your own GGUF models, modes, vocabulary, history, and updates. The only paid thing is Sayly Cloud, a hosted API for people who do not want to manage keys or models.
What runs on my Mac?
Two models. Parakeet v3 turns your speech into text, and S1-mini by Superwhisper cleans it up: fillers out, punctuation in, the tone you asked for. Both run on your Mac.
Which Macs?
macOS 14 or later. Apple silicon is recommended.
Does my audio leave my Mac?
No, unless you choose that. With the local models nothing leaves your Mac. Audio goes out only if you add a cloud provider with your own key or turn on Sayly Cloud, and then only to that provider. Sayly keeps your audio for 14 days by default so you can play a dictation back, then deletes it; you can turn storage off.
Can I use my own API keys?
Yes. Paste a key for an OpenAI-compatible endpoint, Deepgram, or Anthropic, or point Sayly at Ollama on your machine. Sayly talks to the provider directly with your key.
Windows?
Later. Sayly is Mac only today. Windows comes after the Mac app is where it should be.
Privacy
Nothing leaves your Mac by default: the speech model and the cleanup model run on your machine. There is no analytics and no telemetry, in the app or on this site. Your speech reaches a cloud provider only if you add one with your own key, or turn on Sayly Cloud.