Core concepts
Understand how FindIP's semantic patent search engine works under the hood.
What is vector embedding?
Vector embedding turns text into hundreds of dimensions of numbers. Texts with similar meanings end up near each other in vector space, so you can find related documents based on meaning even when keywords differ.
FindIP's embedding pipeline
- Split each patent into sections (title, abstract, claims, description)
- Vectorize each section with the embedding model and store in a vector DB
- Vectorize the query with the same model and find the nearest vectors
- Rerank with a reranking model for precise final ordering
Semantic search
Unlike traditional keyword matching, FindIP uses semantic embeddings to capture the hidden intent and meaning of a query. You can search with natural-language sentences, technical problems, or solution approaches and still find highly relevant patents — even when the wording differs.
Search example
Keyword search: "lithium battery overheating" — only matches the exact words.
Semantic search: "how to prevent thermal runaway in EV battery packs" — understands the technical context and intent.
Hybrid search (2-track system)
To get the best results FindIP combines two search strategies in parallel:
Track B — Hybrid (paragraph level)
Dense vector (semantic) + Sparse (BM25 keyword) search combined. Finds very precise matches at paragraph level.
Track C — Dense (document level)
Pure dense-vector search at document level. Effective for broad topical or wide-ranging technical similarity.
Results from both tracks are fused via the RRF (Reciprocal Rank Fusion) algorithm to produce the final ranking.
Supported countries
FindIP indexes patent data from the world's major patent offices.
| Country code | Office | Language |
|---|---|---|
KR | Korea (KIPO) | Korean |
US | United States (USPTO) | English |
JP | Japan (JPO) | Japanese |
CN | China (CNIPA) | Chinese |
EP | European Patent Office (EPO) | English, French, German |
Patent document structure
Each patent document is composed of sections that can be searched and retrieved individually:
AbstractClaimsDescriptionFiguresMetadataAPI endpoint overview
FindIP exposes the following core API endpoints:
| Method | Endpoint | Description |
|---|---|---|
POST | /search | Run a semantic patent search |
GET | /documents/{id} | Patent detail (Smart Resolution supported) |
GET | /figures/{id}/{filename} | Patent drawing image |
POST | /trends | Natural-language trend / statistics analysis |
POST | /stats | SQL-based aggregation directly on the database |
Base URL: https://api.findip.ai