Privacy Policy

What we process

• The manuscript file (.docx) you upload.
• The target journal homepage URL you provide.
• A pseudonymous account identifier, and basic operational logs (timestamps, IP, action) for security and abuse prevention.

How we protect it

• Encryption at rest: uploaded manuscripts and generated outputs are encrypted in storage.
• Private, expiring downloads: files are served through signed, time-limited links scoped to your job. We never expose raw storage paths.
• Access control: jobs are restricted to their owner; raw storage locations are never returned to the browser.
• Error & prompt logging: operational logs record what happened, not your manuscript text. Where prompts are logged for debugging, manuscript content is hashed, not stored in clear.

Retention & deletion

Manuscripts and outputs are automatically purged after a retention window (30 days by default). You can also delete a manuscript and all of its outputs at any time using the “Delete manuscript & outputs” control on the results page — deletion removes the stored files and database records.

AI providers & training

We do not use your documents to train AI models. To produce a review, manuscript text is sent to the configured large-language-model provider(s) — which may include DeepSeek, OpenRouter-routed models, Mimo, or Anthropic — solely to generate your outputs. We request no-training / no-retention handling from providers where they support it; each provider also publishes its own data-handling terms, which govern their processing. If you cannot send your content to a third-party model, do not upload it.

Confidential & patient data

Do not upload identifiable patient information unless you have authorization and the system is configured for compliant handling. This service is not, by default, a HIPAA-compliant environment.