PQAI
projectpq.ai
What it is
PQAI is an AI-powered patent search platform intended to help find prior art relevant to an invention. Searchers can describe an idea in natural language without special formatting or technical expertise. Its tools include semantic search, CPC Classification Prediction, Technical Concept Extraction, Patent Data Retrieval, and claim analysis. The Free plan searches across 100M+ patents and returns 20 results per query at standard speed. PQAI+ offers 500 results per query, faster search, full-text excerpts, and export. The service is available on the web and through an API, and PQAI describes it as open source. API plans include MCP server access and compatibility with Claude, ChatGPT, and Gemini. PQAI also provides an iFrame embed with instructions for WordPress, Wix, and custom-built websites. The company says it erases each query from its servers after generating results. API requests use token-based authentication and HTTPS, alongside rate limiting and access controls. PQAI identifies individual inventors, law firms, patent examiners, enterprises, and R&D teams as users. Paid plans start at $20 a month; commercial integration into another product requires a commercial API or Enterprise plan.
Spec sheet
- What it does
- PQAI uses AI to search patents and technical literature for relevant prior art from plain-language invention descriptions.Read 15 Oct 2026 · github.com
- Purpose
- PQAI provides AI-powered patent searches to help users find relevant prior art.Read 7 Oct 2026 · projectpq.ai
- Audience
- PQAI identifies individual inventors, law firms, patent examiners, enterprises, and R&D teams as users.Read 7 Oct 2026 · projectpq.ai
- Integrations
- The API plans include MCP server access and state compatibility with Claude, ChatGPT, and Gemini.Read 7 Oct 2026 · projectpq.ai
- Security
- PQAI says API requests use token-based authentication and HTTPS, with rate limiting and access controls.Read 7 Oct 2026 · projectpq.ai
- Privacy
- PQAI says it does not store search queries and erases each query from its servers after generating results.Read 7 Oct 2026 · projectpq.ai
- Support
- Free and PQAI+ users can contact support through Help Docs and a contact form, with a stated aim to respond within one business day.Read 7 Oct 2026 · projectpq.ai
- API authentication
- API requests use unique token authentication and HTTPS transport.Read 15 Oct 2026 · projectpq.ai
- API documentation
- The public API guide documents routes, path/query parameters, authentication and example responses.Read 15 Oct 2026 · api.projectpq.ai
- API limit
- The PQAI+ plans include 20 API requests per month, and the page says responses stop until the next billing cycle after the limit is reached.Read 7 Oct 2026 · projectpq.ai
- Browser storage
- Saved or bookmarked results are stored in the user’s browser rather than uploaded to PQAI servers.Read 15 Oct 2026 · projectpq.ai
- Commercial API use
- Commercial integration into a user's own product or platform requires a commercial API or Enterprise plan.Read 7 Oct 2026 · projectpq.ai
- Company origin
- LinkedIn says AT&T initiated the PQAI project and that other groups and companies support it.Read 15 Oct 2026 · linkedin.com
- Company profile
- LinkedIn lists PQAI as a nonprofit software-development organization headquartered in the United States with 2–10 employees.Read 15 Oct 2026 · linkedin.com
- Concept extraction
- The API and search product extract technical concepts from supplied text or patents.Read 15 Oct 2026 · projectpq.ai
- Coverage expansion
- A December 17, 2024 update says coverage reached patents from 68 countries and more than 100 million research papers.Read 15 Oct 2026 · search.projectpq.ai
- CPC prediction
- PQAI predicts Cooperative Patent Classification codes from technical text.Read 15 Oct 2026 · projectpq.ai
- Export
- Search results can be exported for offline review, with the paid search plan allowing up to 500 results.Read 15 Oct 2026 · projectpq.ai
- Funding fact
- The same IFIA report said AT&T funded the PQAI project with $2 million.Read 15 Oct 2026 · ifia.com
- Independent availability claim
- Founders Legal describes its PQAI collaboration as offering unlimited free searches, confidential processing and downloadable reports.Read 15 Oct 2026 · founderslegal.com
- Independent launch news
- IFIA reported on May 13, 2021 that PQAI was an open-source nonprofit initiative launched with IFIA involvement.Read 15 Oct 2026 · ifia.com
- Independent result-limit complaint
- Founders Legal’s embedded PQAI tool explains that its interface intentionally presents only ten results and directs users needing broader coverage to professional services.Read 15 Oct 2026 · founderslegal.com
- Licence
- The public PQAI repository states that the project is open source under the MIT licence.Read 15 Oct 2026 · github.com
- Limit: cross-device saved results
- A G2 reviewer notes that browser-stored results are inconvenient when switching devices.Read 15 Oct 2026 · g2.com
- Limit: interface polish
- A G2 reviewer says the interface can feel basic or in need of refinement compared with premium commercial tools.Read 15 Oct 2026 · g2.com
- Local deployment
- The repository points developers to wiki instructions for deploying PQAI locally for experimentation or development.Read 15 Oct 2026 · github.com
- MCP integration
- The paid API plans include access to an MCP server that works with Claude, ChatGPT and Gemini.Read 15 Oct 2026 · projectpq.ai
- Natural-language search
- Users can describe an invention in ordinary language without special query formatting or patent-search expertise.Read 15 Oct 2026 · projectpq.ai
- Non-patent literature
- The search platform includes scientific and other non-patent literature alongside patents.Read 15 Oct 2026 · search.projectpq.ai
- Notable update
- The public updates page records an April 7, 2025 release adding country filters, NPL-only search and stability fixes.Read 15 Oct 2026 · search.projectpq.ai
- Open source
- PQAI describes itself as an open-source platform and invites the developer community to improve it.Read 7 Oct 2026 · projectpq.ai
- Patent coverage
- The free plan includes search across 100M+ patents.Read 7 Oct 2026 · projectpq.ai
- Patent data retrieval
- Structured patent data such as titles, abstracts, claims and citations can be retrieved through the API.Read 15 Oct 2026 · projectpq.ai
- Praise: API integration
- G2 reviewers praise the API as straightforward, well documented and easy to connect to internal tools.Read 15 Oct 2026 · g2.com
- Praise: natural-language usability
- G2 reviewers repeatedly praise entering plain-English invention descriptions instead of constructing Boolean searches.Read 15 Oct 2026 · g2.com
- Praise: relevance and speed
- G2 reviewers repeatedly describe results as highly relevant and returned quickly, reducing manual prior-art work.Read 15 Oct 2026 · g2.com
- Privacy claim
- PQAI says it does not log or retain search queries and erases a query after generating results.Read 15 Oct 2026 · projectpq.ai
- Rating on G2
- 4.9 out of 5 from 8 reviewsRead 15 Oct 2026 · g2.com
- Search features
- The platform includes CPC Classification Prediction, Technical Concept Extraction, and Patent Data Retrieval.Read 7 Oct 2026 · projectpq.ai
- Search input
- Users can describe an invention in natural language without special formatting or technical expertise.Read 7 Oct 2026 · projectpq.ai
- Self-hosting
- Private Server is offered for on-premise or private-cloud deployment so queries remain inside the customer environment.Read 15 Oct 2026 · projectpq.ai
- Semantic matching
- The engine ranks semantically relevant prior art rather than relying only on keyword hits.Read 15 Oct 2026 · projectpq.ai
- Support model
- Repository support is handled through GitHub issues, while paid users receive helpdesk/email support and enterprise customers get priority assistance.Read 15 Oct 2026 · github.com
- Website embedding
- PQAI supplies an iframe embed URL and instructions for WordPress, Wix and custom websites.Read 15 Oct 2026 · projectpq.ai
Plans and prices
Strong points and limits
Strong points
- Natural-language invention descriptions need no special format.
- Free plan searches across 100M+ patents.
- Includes CPC prediction and technical concept extraction.
- API plans include MCP server access.
- Queries are erased after results are generated, according to PQAI.
Limits
- Free plan returns 20 results per query.
- PQAI+ plans include 20 API requests per month.
- Commercial product integration requires a commercial API or Enterprise plan.
Who it suits
PQAI suits individual inventors, law firms, patent examiners, enterprises, and R&D teams searching for prior art. Its API and website embedding options also suit developers adding patent search to a product or site.
Verdict
PQAI offers natural-language patent searching with named analysis tools and a free plan covering 100M+ patents. Paid plans start at $20 a month, while commercial integration requires a commercial API or Enterprise plan.
Where Specifiction places it
- AI Patent Search Software#2 of 248.7