Privacy audit
Most apps claim to be private. This one lets you verify it. A live monitor records every network request your browser makes and classifies it by destination — so you can see, in real time, that no document is uploaded to a third party. Below it: the formal threat model and exactly where each feature processes your data.
See the privacy-aware router that decides what may leave your device →No third-party uploads detected
Every request this session went to the app's own origin or a service you configured — no document was sent to an unknown host.
Live network monitor
Watching… Convert a file or run a tool in another tab and the requests appear here in real time.
Where each feature processes your data
FFmpeg WebAssembly — the file never leaves your device.
Processed by the app's own server, in memory, then discarded.
Python backend on the same deployment; no third-party service.
TextRank runs entirely client-side with no network call.
Deterministic scoring on the app's backend; résumé discarded after scoring.
Embeddings computed on the backend; no document leaves the deployment.
ONLY this opt-in action sends the selected text to your configured AI provider.
Threat model — who can see your document
Media never transmitted; other traffic is TLS to first-party origin only.
No document is sent to an AI provider unless you explicitly opt in to an AI action.
No third-party scripts; CSP + a live network monitor prove zero external calls.
Files are processed in memory and discarded; the database stores metadata only.
COOP/COEP isolation limits cross-origin access, but an extension with page access is outside the app's control.
Anything running on your own machine is outside any web application's security model.
This dashboard is itself deterministic: the verdict is computed from real intercepted requests, not a promise.