Privacy
Your conversations, under your control.
Speech to text runs locally on your Mac. For intelligence you pick the LLM, and that can be a local one too.
How your data flows
Three steps, you decide the middle one.
- 01Local
Speech to text
A speech model runs on your Mac and turns your mic and call audio into a transcript. Audio is never uploaded.
- 02Your choice
Intelligence
Notes and suggested questions come from an LLM. You decide which one, and where it runs.
- 03Local
Storage
Transcripts and notes are stored on your Mac. Delete them there and they are gone.
Two ways to run
Cloud LLM or fully local. Your call.
Cloud LLM
Add your own API key in Settings. Only the text you choose to send goes to that provider, under their privacy terms. Roofle never sees it.
Local LLM
Point Roofle at a model running on your own machine. Speech to text and intelligence both stay local, so nothing leaves your Mac.
The local LLM option works for notes and suggested questions, so you can run Roofle without any data leaving your machine.
Our promise
What we never do.
No bots
Nobody extra joins your call. Roofle runs from your menu bar.
No recordings on our side
We never receive your audio, transcripts or meeting content.
Open source
The code is public, so you can check what Roofle does with your data.
Consent rules vary by country and company. Let meeting participants know you are transcribing.