The current AI Doc Scan implementation does not upload your selected PDF or extracted document text to a marcoderspace server.
PDF.js reads the document in your browser. Explainable diagnostics run in JavaScript on the device. When deep scan is enabled, the browser downloads an AI-text classification model and performs inference locally.
The application currently loads software/model resources from third-party CDNs and Hugging Face. Those providers may receive ordinary web-request metadata such as IP address, user agent and request time when resources are fetched. The PDF itself is not sent as part of those requests by AI Doc Scan.
The scanner currently requires no user account and stores no uploaded document on marcoderspace infrastructure.
If advertising or analytics are introduced, this privacy notice will be updated before those services are enabled. Where required, an appropriate consent-management mechanism will be used.