Skip to content

Privacy first

Your paper never leaves your device

ReviewerAI runs the paper-review model locally in your browser. The manuscript is not uploaded for review unless you choose a cloud-based action such as saving it to an account. That is the whole privacy promise in one sentence.

The model comes to you — your data does not go to the cloud

Most AI tools work the other way: you send your document to their servers and they send an answer back. ReviewerAI flips this. The review software is downloaded to your computer once, and then your paper is reviewed where it already is.

  • Your PDF is opened and read inside your own browser tab.
  • The checks run on your computer using your own hardware.
  • Findings are shown to you on the same device — nothing needs to be sent away to produce them.
  • Closing the tab without saving leaves no copy of your paper on our systems.

What stays local, and what does not

What stays on your device and what is sent
ActionLeaves your device?Explanation
Running a reviewNoThe PDF and the findings stay in your browser.
Downloading the review softwareOnly normal downloadsYour browser fetches the app files, like any website. Your paper is not part of this.
Saving to your account (optional)Yes — only if you click saveThe saved copy lets you reopen the review later. Skip it and nothing is stored.
Contacting supportYes — only what you sendIf you email us, we see that email. We never pull papers from your browser.

Why this matters for universities

Many universities restrict sending unpublished research, student work, or grant drafts to outside AI services — and the rules differ between the public version of a tool and a university-provided one. A local review sidesteps the whole question: there is no transfer to approve, log, or worry about.

  • No need to ask which endpoint to use — the review does not call an endpoint.
  • No training use — your paper is never used to train future models.
  • No third-party readers — no staff, contractors, or data pipelines see the manuscript.
  • Supervisors can recommend it to students without clearing a new cloud vendor.

Review confidential work with a clear conscience

Thesis chapters, grant proposals, and unpublished manuscripts can all be checked locally before anyone else sees them.