What leaves your machine.

Short version

Scribewave Airgap is designed to run entirely on your own device. Audio, video, transcripts, and embeddings are written to your local app-data directory and never uploaded. We do not run analytics, ship telemetry, or collect any usage data by default.

The complete network inventory

The app makes outbound calls only in these cases:

• License activation — one HTTPS POST to api.airgap.scribewave.com when you enter your license key, exchanging the key for a device-bound activation token. Includes a stable hash of your device and your license key. No audio, no filenames.
• Update checks (opt-in) — if you enable "Check for updates," the app fetches a small JSON manifest from the same domain on launch. You can disable this in Settings → Updates.
• BYO LLM keys (opt-in) — if you configure Anthropic, OpenAI, Gemini, or Ollama for AI summaries, the relevant prompts are sent to your chosen provider when you press the AI button. Scribewave never sees the request.
• Model downloads — Whisper, diarization, and embedding model files are downloaded on first run from public mirrors. After that, models load from disk and the app makes no further model traffic.

What we keep on our servers

On our backend (the license API), we store the email you purchased with, your license key, an opaque hash of each activated device, and the timestamps shown in your portal. We never receive your recordings, transcripts, or any content from inside the app.

Cookies and analytics

This marketing site does not use cookies, fingerprinting, or third-party analytics. The session JWT for the license portal is stored in your browser's localStorage and is cleared when you sign out.

Contact

Questions, deletion requests, or compliance inquiries — hello@scribewave.com. We respond to every message.