Your book is
not our content.

What happens to an upload

1. The source file is placed in private temporary object storage.
2. A background process extracts structured text and validates it.
3. After successful extraction, the original DOCX, EPUB, or TXT file is deleted.
4. The extracted content is retained for the configured scan window, currently 24 hours, then automatically deleted.

Model processing

To perform a scan, manuscript text is sent server-to-server to a configured third-party model through OpenRouter. We request zero-data-retention routing and deny provider data collection where supported. Provider handling is governed by OpenRouter and the selected model provider; we do not make claims beyond those controls.

Training

We do not use uploaded manuscripts to train our own models. This service is a manuscript audit, not a data-collection program.

Reports and payments

Reports and the minimal metadata needed to operate the service are retained for the configured report period, currently 30 days. Stripe processes payment details; this application stores only relevant session and payment identifiers, amount, currency, status, and the customer email needed for access or a refund request. Full card details never pass through this application.

Abuse prevention

Free scans use Cloudflare Turnstile and pseudonymous, keyed hashes derived from IP and browser-device identifiers. Raw IP addresses are not stored in the application database. These records are used only to enforce reasonable scan limits.

Your controls

The secure report page includes a delete control. It removes manuscript content, findings, and associated model-request records where technically possible. Minimal payment records may remain where needed for accounting, disputes, and fraud prevention.

Access links

The report URL contains a high-entropy secret. The database stores only a keyed hash of that secret. Treat the URL like a private magic link and do not forward it to people who should not see the report.