Asclevor Technology
The infrastructure behind medical intelligence.
Asclevor turns published clinical knowledge into structured, retrievable infrastructure — embeddings, semantic retrieval, and ranking over a growing index of patient cases.
- Cases processed
- 167,000+
- Cases indexed
- ~40,000
- Embedding model
- BAAI/bge-small-en-v1.5
- Access
- Open — no key required
Core technology
Five layers. One foundation.
AI embeddings
Queries and cases are embedded with BAAI/bge-small-en-v1.5, a high-dimensional embedding model.
Semantic retrieval
Similarity search across the indexed case base follows meaning, not keywords.
Medical knowledge
A growing index of published patient cases, each traceable to its source PMID.
Relevance ranking
Results are ordered by semantic similarity, with a transparent score on every case.
Data infrastructure
More than 167,000 published cases processed — approximately 40,000 in the current index.
Architecture
How a query becomes results.
- Query
A clinical question, in natural language.
- Query Processing
The query is embedded with BAAI/bge-small-en-v1.5.
- Semantic Retrieval
Similarity search across the indexed case base.
- Ranking
Cases ordered by semantic similarity score.
- Results
Structured JSON with scores and source PMIDs.
Reliability & security
What you can count on today.
- Encrypted transport
- All traffic to and from the platform is encrypted in transit via TLS.
- EU infrastructure
- Infrastructure is hosted within the European Union, with data residency by design.
- Data handling
- No patient data, ever. The platform serves published medical knowledge and account data — nothing else.
- Open access
- The API is currently open — no key is required to make a first request, and results are metered fairly for everyone.
- Monitoring Planned
- Usage analytics, error tracking, and a public status page are on the roadmap, alongside ISO 27001 certification.
Continue