Deep Review is an AI system from Google Research that checks long technical documents, such as research papers, design documents, presentations and legal or financial documents, for logical gaps, mathematical errors, internal inconsistencies and unsupported claims. Google released it publicly on October 14, 2026. It grew out of Google’s experimental Paper Assistant Tool, which gave opt-in feedback to the authors of more than 16,000 manuscripts submitted to the STOC, ICML, NeurIPS and ICLR computer science conferences.
It is free for light use: every user gets two complimentary document reviews per week, and people who need more can connect a Gemini API key for higher limits and programmatic access through the Deep Review API. At launch it accepts PDFs and plain text, with multi-file input planned. Google says uploaded documents are never used to train its models and are stored for one year only so they can be shown back to the user. The system splits a document into segments, gives harder sections such as proofs more computing effort, reviews the segments in parallel with separate AI agents, and then combines and cross-checks the findings.
Source: Google Research, “Analyzing technical documents with Deep Review” (October 14, 2026).
Image: “Example of copyedited manuscript” by Phoebe, CC BY-SA 3.0, via Wikimedia Commons.