paperRR
Start analyzing

About the product

About paperRR

paperRR analyzes AI-writing signals in long-form text and makes the supporting evidence reviewable.

Last updated:

What problem do we solve?

A single score cannot support a consequential decision. paperRR links each finding to passages and evidence in the source text.

  1. 01Overall findingInterpret the finding by the share of evaluated text
  2. 02Segment scoresNarrow down passages for review
  3. 03Five-part score breakdownShows how the five contributions add up
  4. 04Word-level evidenceReturn to specific locations in the original text

What don't we do?

paperRR does not identify authors or reconstruct writing processes. Its reports support contextual review; they are not verdicts.

How should I interpret the published metrics?

The home page identifies the tested model, sample size, accuracy, and training-corpus scale. Accuracy reflects that test set, not every text. Historical results are labeled, and untested scenarios remain blank. Review recall, false-positive rates, protocols, sample sizes, and limits under Methodology and results.

How do I verify a report ID?

New web reports and PDFs use the same server-issued ID. After receiving a report, use Report ID verification to retrieve the issuance summary and compare each item with the report.

  1. Copy the complete ID from the PDF cover or the top of the web report, preserving capitalization and symbols.
  2. Paste the ID on the verification page and complete the CAPTCHA. Opening a link that contains an ID does not automatically run a search.
  3. Compare the issuance date, original length, AI-text share, language, document type, and model version.

Verification confirms the issuance record and its summary. It does not determine whether a PDF, screenshot, or report body has been altered, and it does not prove that the analysis is correct. Legacy reports without an ID cannot be verified online. If no record is found, check the ID and contact the report provider; a failed lookup alone does not prove that a file is forged.

Anyone with the complete ID who passes the CAPTCHA can view the summary but cannot retrieve submitted text, segment-level evidence, or the full report. Before sharing an ID, read the Privacy for the scope of public lookup and retention periods.

Contact

For product, policy, or evaluation questions, email pattinghead@outlook.com.