Follow the evidence
Every record points to a first-party source, with a retrieval date and content fingerprint.
Find research, models, benchmarks, and software for web, desktop, and mobile agents. Filter by task or platform, then inspect the primary sources.
Clone for local search and RAG exports. Records retain source evidence, review states, and scoped licenses.
Browse download formats ↗git clone https://github.com/reacher-z/Awesome-CUA.gitcd Awesome-CUA
python3 scripts/catalog.py search "grounding" --kind model --status curated --jsonPython 3.11+; no package installation or account needed.
You can browse the library or download the data while it loads.
Check your connection and try again.
Try a broader term or clear a filter. Many papers have not yet been assigned platform or topic tags.
Reviewed: relevance and annotations checked against a primary source. Metadata only: bibliography verified; relevance pending. Neither status means experiments were reproduced.
See which records have source review, which have metadata only, and which identifiers remain unresolved.
Every record points to a first-party source, with a retrieval date and content fingerprint.
Papers are not projects. Code licenses are not data licenses. A model checkpoint is not a model family.
Comprehensive coverage is our goal. The unreviewed queue and source boundaries stay visible along the way.
Found a paper, dataset, tool, or a detail we missed? Add your knowledge to the computer-use community.
Browse source-backed pages directly, with or without JavaScript. Reviewed entries include concise annotations, provenance, and separate license declarations.
Understand computer-use agents, choose a benchmark for the capability you need, or find training data with the right supervision and license.