Privacy Policy

Last updated: February 7, 2026

The short version

We collect nothing. Push to Transcribe processes all audio locally on your device (Mac or Windows PC). No audio, text, or usage data is ever sent to any server. We have no servers to send it to.

Data Collection

Push to Transcribe does not collect, store, or transmit:

• Audio recordings
• Transcribed text
• Usage analytics or telemetry
• Device identifiers
• Personal information
• Crash reports
• Any data whatsoever

Local Processing

All transcription happens entirely on your device using local AI models (whisper.cpp, plus Apple Speech on Mac). Your audio never leaves your device. There is no cloud component, no API calls, and no network requests related to transcription.

Network Access

Push to Transcribe requires network access only for:

• Downloading AI models — On first run, the on-device speech models (Whisper) and a voice-activity model (Silero VAD) are downloaded once from Hugging Face, their official source. These are plain file downloads; no personal data is sent.
• License activation — When you activate, the app validates your license with Polar (polar.sh), our payments and licensing provider. It sends your license key and an activation label that includes your computer's name. No audio or transcribed text is ever sent.
• Software updates — The app checks our own servers (Cloudflare) for new versions: on macOS via Sparkle, on Windows via Velopack. This is a standard update check. Push to Transcribe is distributed as a direct download, not through the Mac App Store or Microsoft Store.
• Sending feedback (macOS only) — If you choose to submit a bug report or feature request from inside the Mac app, the text you type is sent to our feedback endpoint (a Cloudflare Worker). This happens only when you actively submit feedback, and only sends what you wrote — no audio, no transcripts, nothing automatic. The Windows app has no in-app feedback feature.

After initial model download, Push to Transcribe works fully offline.

Data Storage

Your transcription history, word replacements, and preferences are stored locally in your app container on your device, encrypted at rest (on Windows via the operating system's DPAPI). This data is never synced, uploaded, or shared. You can delete all local data at any time by removing the app.

Third-Party Services

Push to Transcribe does not integrate with any analytics, advertising, or tracking services, and your audio and transcribed text are never sent anywhere. The only third parties the app ever contacts are: Cloudflare, which serves the installer and update feed (our download endpoint logs standard request data such as IP address and browser user-agent, like any web server, and on macOS only it also hosts the endpoint that receives any in-app feedback you choose to submit); Polar (polar.sh), which processes your purchase and validates your license (it receives your license key and an activation label that includes your computer's name); and Hugging Face, which hosts the AI models you download once on first run. The Windows installer is code-signed via Azure Trusted Signing at build time. Push to Transcribe is not distributed through the Mac App Store or Microsoft Store.

No Account Required

Push to Transcribe does not require or support user accounts. There is no registration, login, or authentication of any kind.

Children's Privacy

Since we collect no data from anyone, we also collect no data from children.

Changes to This Policy

If we ever update this privacy policy, changes will be posted on this page with an updated date. Our commitment to collecting zero data is unlikely to change.

Contact

Questions about this privacy policy? Email us at [email protected].