Query a Knowledge base

POST /v1/knowledge-bases/{knowledge_base_id}/query

Runs hybrid retrieval over one version of the Knowledge base, the active one unless version names another or working_set, and answers the passages with their citations (1.53.0; knowledge-bases.md section 6). Semantic and keyword candidates, each restricted to the version's manifest and the filter, are fused by reciprocal rank fusion, then dropped below min_score, cut at top_k and packed within max_context_tokens. Without a content class in include the passages carry no text and the citations no names, and nothing is audited; naming text or names makes the query an audited content read. No version named and none active is 422 validation_failed with knowledge_base_no_active_version; a version that does not exist is 404 not_found, and a retired one 409 knowledge_base_version_retired. An archived Knowledge base still answers. From 1.54.0 the api runs no model, so it asks a worker of the ingest role for the query's vector through a query embedding task (KnowledgeQueryTask), holding the query sealed in that task for the wait only, and answers 503 dependency_unavailable when no vector comes within 5 seconds (a target) or the worker answers embedding_failed; elapsed_ms includes the wait.

Request

Set VOCAL_API to the base URL of your API and VOCAL_API_KEY to an API key with the knowledge:read scope (Authentication and keys). Replace each placeholder, such as agt_... or <name>, with your own value.

curl -X POST "$VOCAL_API/v1/knowledge-bases/kb_.../query" \
  -H "Authorization: Bearer $VOCAL_API_KEY" \
  -H "Idempotency-Key: $(uuidgen)" \
  -H "Content-Type: application/json" \
  -d '{ "query": "<query>" }'

const response = await fetch(`${process.env.VOCAL_API}/v1/knowledge-bases/kb_.../query`, {
  method: 'POST',
  headers: {
    Authorization: `Bearer ${process.env.VOCAL_API_KEY}`,
    'Idempotency-Key': crypto.randomUUID(),
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({
    "query": "<query>"
  }),
});
console.log(response.status, await response.text());

import os
import uuid

import requests

response = requests.post(
    f"{os.environ['VOCAL_API']}/v1/knowledge-bases/kb_.../query",
    headers={
        "Authorization": f"Bearer {os.environ['VOCAL_API_KEY']}",
        "Idempotency-Key": str(uuid.uuid4()),
    },
    json={
        "query": "<query>",
    },
)
print(response.status_code, response.text)

Parameters

Request body

Content type application/json, required: KnowledgeQueryRequest.

Responses

Fields of the 200 answer

The shape is KnowledgeQueryResult.