Asclevor Developers

Build with Asclevor.

One open API over Asclevor's clinical retrieval infrastructure. Send a clinical question as JSON — get ranked published cases with similarity scores and source PMIDs.

curl -X POST https://api.asclevor.com/v1/search \
  -H "Content-Type: application/json" \
  -d '{
    "query": "fever, productive cough, and confusion in an older adult",
    "limit": 5
  }'
The API is open — no key required.

One endpoint. Structured answers.

Everything runs through a single search endpoint. Send a clinical question in natural language; receive the most similar published cases, ranked and fully sourced.

Endpoint
POST /v1/search
Authentication
None — open access
Request format
JSON
Response format
JSON · scores · PMIDs

Capabilities

What you get out of the box.

  • Clinical search

    Natural-language queries over published patient cases.

  • Semantic retrieval

    Meaning-based matching, not keyword matching.

  • Structured results

    Predictable JSON — no scraping, no parsing PDFs.

  • Transparent ranking

    A similarity score on every result.

  • Source traceability

    Every case carries its PubMed identifier.

  • Open access

    No API key required while Asclevor is in active development.

{
  "query": "fever, productive cough, and confusion in an older adult",
  "model": "bge-small-en-v1.5",
  "count": 3,
  "results": [
    { "pmid": "31784915", "score": 0.93 },
    { "pmid": "33748325", "score": 0.88 },
    { "pmid": "32409471", "score": 0.84 }
  ]
}

Anatomy of a response.

Every field is predictable — build against it once and it stays built.

results
Published cases ranked most-similar-first.
results[].score
Semantic similarity to the query — higher means more similar.
results[].pmid
PubMed identifier for the published case, for direct citation.
model
The embedding model used for retrieval.
count
How many cases came back.

Resources

Developers

Start building today.

The API is open — no key required. Tell us what you're building.