About AwesomeCUA and its editorial policy
AwesomeCUA is a source-backed catalog of resources for agents that operate computer interfaces. Its purpose is to make research and engineering assets easier to find, compare, and trace to their original publishers.
Who maintains the catalog?
The AwesomeCUA project publishes the catalog through its GitHub organization. Annotations and taxonomy are produced with AI-assisted editorial review. This is not a claim of independent peer review, benchmark reproduction, or affiliation with listed publishers.
What is included?
GUI and hybrid GUI/code agents, visual grounding, web and mobile navigation, training data, evaluation environments, execution infrastructure, safety research, and relevant products. Supporting tools may be marked adjacent.
How are claims checked?
Community lists and semantic searches discover candidates. Retained arXiv and publication sources support bibliographic facts. Official papers, repositories, documentation, and model/dataset cards support curated annotations. Every record retains source URLs, retrieval timestamps, and content hashes. Source refreshes do not silently inherit editorial approval.
What do review states mean?
- Curated: a primary source was inspected for identity, scope, and the concise annotation.
- Metadata verified: bibliographic fields were checked; relevance, taxonomy, and findings are still pending.
- Candidate: an identifier was discovered but is not a published catalog record.
Licenses and corrections
Code, data, weights, and document licenses are separate. Public availability does not establish reuse rights. Original catalog annotations and code use MIT; upstream resources keep their own terms. Report errors through a source-backed correction.
Indexing policy
Reviewed resource pages and useful collection pages are eligible for indexing. Metadata-only resource pages remain accessible, with noindex,follow until their relevance is reviewed. The full bibliography remains searchable and downloadable.