Use Case

Your voice never
leaves your Mac.

Most tools say "your privacy is important to us." We say: your audio can't reach us. Those are different statements.

Model runs on-device

Whisper and Parakeet are downloaded once and execute in memory on your Mac. Your audio is processed by your own hardware, not a remote server.

No network calls during transcription

During a transcription session, Tellaflow makes zero outbound requests. You can verify this with Little Snitch, proxyman, or any network monitor.

Architecturally private

We didn't build a server that receives your audio and then promise to delete it. We didn't build a server at all. Different thing.

What we don't do

We don't receive your audio files
We don't receive transcriptions or translations
We don't know what you said, what app you used, or when
We don't have a server endpoint that receives voice data
We don't log usage events tied to your content
We can't comply with a subpoena for your audio — it's not here

We do collect anonymous crash reports if you opt in during onboarding. These contain stack traces and device specs — never audio, never transcriptions. You can disable this at any time in Settings.

Ready to stop typing?

Free, forever. No account. No API key. Just your Mac and your voice.

Download Tellaflow free
Apple Silicon M1+ · macOS 13+ · 8 GB RAM