Contract Understanding Atticus Dataset (CUAD)

Contract Understanding Atticus Dataset (CUAD)(Commercial Contracts) v1 contains 13,000+ labels in 510 commercial legal contracts with rich expert annotations. The dataset has been manually labeled under the supervision of experienced lawyers to identify 41 types of legal clauses in commercial contracts that are considered important in contract review in connection with a corporate transaction, including mergers & acquisitions, corporate finance, investments, initial public offering, etc.

CUAD is the first dataset in The Atticus Project's open-source research portfolio, which has since grown to include MAUD (merger agreement understanding, accepted at EMNLP 2023) and ACORD (clause retrieval, accepted at ACL 2025) — together representing one of the largest expert-annotated legal AI benchmarks in existence.

This Handbook contains selected guidelines developed by The Atticus Project to guide our volunteer annotators in curating CUAD (Commercial Contracts). This Handbook contains our interpretation of selected Atticus Labels and contract clauses solely for the purpose of curating CUAD and does not constitute legal advice.

This Handbook is subject to the disclaimers set forth at atticusprojectai.org/disclaimer, and the processing, storage and sharing of your personal data and your related rights is subject to the privacy policy set forth at atticusprojectai.org/privacy-policy.

Access to CUAD on our website is free. The full Handbook is available for purchase as a downloadable document. We encourage you to share your suggestions and ideas by emailing info@atticusprojectai.org.

This Handbook and our other work products are created by volunteers of The Atticus Project. Support our continuing efforts to provide high-quality free content by purchasing a downloadable version of the full Handbook.