Route reference
Semantic concept search
GET/v1/search/conceptssearch:read
Finds concept candidates from loose wording, questions, themes, or partial ideas. Use it for discovery.
When to use it
- A user types a theme or problem rather than an exact concept name.
- You need ranked candidates for a search UI, assistant, or recommendation step.
- You plan to call the concept detail route after a user selects a result.
Request
- query
- Required search text.
- top_k
- Optional result count, 1 through 50. Defaults to 10.
- field_id_filter
- Optional field ID used to narrow concept results.
Response
- items[]
- Concept candidates with id, canonical_name, definition, field IDs, and score.
- score
- Retrieval relevance for ranking, not truth or importance.
- meta.vector_search
- Whether semantic matching was available for this request.
How output is produced
- Compares the query against the public concept search surface using semantic and text signals when available.
- Narrows by field when field_id_filter is supplied.
- Projects each result into a compact concept shape suitable for selection UIs.
Trust and caution
- Use this for discovery, not as a final claim that a concept applies.
- If semantic matching is unavailable for a request, results may rely more on text overlap.
- A narrow field filter can hide useful adjacent concepts.
Request example
curl "https://api.agentecology.com/v1/search/concepts?query=adaptive%20coordination&top_k=5" \
-H "X-API-Key: ae_live_YOUR_KEY" \
-H "Accept: application/json" \
-H "User-Agent: your-product/1.0"Response example
{
"items": [
{
"id": "concept_adaptive_coordination",
"canonical_name": "Adaptive coordination",
"definition": "Coordination that changes as local conditions change.",
"field_ids": ["field_organizational_design"],
"field_id": "field_organizational_design",
"score": 0.82
}
],
"count": 1,
"source": "concepts",
"meta": {
"request_id": "req_...",
"timestamp": "2026-05-26T17:30:00Z",
"vector_search": true,
"metered": true
}
}