Skip to main content

Knowledge API

Manage documents in the RAG (Retrieval-Augmented Generation) knowledge base.

List Files​

GET /v1/knowledge?agent_id=optional_filter

Scope required: files

Response​

{
  "files": [
    {
      "id": "doc_faq",
      "filename": "faq.txt",
      "content_summary": "Company FAQ document",
      "status": "ready",
      "size_bytes": 4521,
      "agent_id": "agent_soporte",
      "uploaded_at": "2026-05-01T10:00:00Z"
    }
  ]
}

Ingest Document​

POST /v1/knowledge/ingest

Upload text content to an agent's knowledge base. The document is queued for embedding and vectorization.

Request​

{
  "agent_id": "agent_soporte",
  "filename": "returns_policy.txt",
  "content": "Our return policy allows returns within 30 days of purchase..."
}

Response​

{
  "file_id": "doc-1715100000000-api-a1b2c3d4",
  "status": "pending",
  "message": "Document queued for RAG ingestion. Use GET /v1/knowledge to check status."
}
tip

The ingestion pipeline runs asynchronously. Check the file status by polling GET /v1/knowledge. Status transitions: pending → processing → ready.