Designed to minimize retained data.
anonymise.online processes submitted text and PDFs for the active workflow and does not send them to an external AI provider.
Text processing
Your source text and restore-map upload are processed to produce the requested result. They are not used to train an AI model and are not stored as a searchable user-document library.
PDF processing
Uploaded PDFs are parsed from an in-memory multipart stream under the request-size limit. PDF bytes are not written to a server-side temporary file, database, Redis, or session state. Page previews and extracted text are returned only to the requesting browser for review.
No external AI detection
Sensitive-data detection uses application-local patterns and offline English and Italian language analysis. The source text is not forwarded to ChatGPT, Claude, Gemini or another hosted AI service.
Temporary workflow state
The review step uses short-lived protected state so the application can carry the anonymized result into finalization without creating a permanent user account.
Premium access and payments
Stripe processes payment information. anonymise.online receives payment status and the checkout email needed to deliver access credentials. It does not receive full card details.
Operational statistics
The administrator can view aggregate request totals, page views, anonymization counts, character volume and paid-plan status. Submitted text, restore maps and full access codes are not included in those statistics.
Your restore maps
Text and PDF restore-map JSON files contain original sensitive values. Store them separately from anonymized output and protect them like secrets. PDF maps are generated in the browser and do not contain a PDF hash or source-file fingerprint.
See the full private AI workflow and the security overview.