PQAI
About this tool
An open-source AI prior-art search server trained on historical patent-examination records. Describe an invention in natural language to find related patents and technical literature. REST APIs cover single-document and combined-document searches, patent-date searches, similarity, element mappings and CPC suggestions. Self-host with Docker or use the hosted API.
- Feature 1: /search/102/ and /search/103/ search individual and combined prior art.
- Feature 2: /prior-art/patent/ finds documents published before a patent's filing date.
- Feature 3: /mappings/ maps query elements to documents.
- Feature 4: /suggest/cpcs suggests CPC classes.
- Feature 5: Dockerfile and docker-compose deployment files are included.
Resource information
- Maintainer
- PQAI project team (pqaidevteam); open-source community project.
- Type
- Tool / library
- Category
- Patents & intellectual property
- Runs on
- Tool or library
- Who it's for
- Researchers and developers screening invention ideas or studying patent retrieval, and technology-transfer teams hosting prior-art search.
- License at time of listing
- MIT
- Source
- GitHub
