FruxonDocs
IntegrationsAssets

Lists chunks for a document using cursor-based pagination.

Returns the ordered chunks produced by the ingestion pipeline — the indivisible units that searches return and that the embedding model encoded. Chunk size and overlap are determined by the asset's embedding configuration, not by this call. Primarily used to inspect what the model actually sees for a given document when debugging poor search relevance or unexpected agent answers.

GET
/v1/tenants/{tenant}/assets/{asset}/documents/{document}/chunks

Authorization

Bearer
AuthorizationBearer <token>

JWT Authorization header using the Bearer scheme. Enter 'Bearer' [space] and then your token.

In: header

Path Parameters

asset*string

The asset ID.

Formatuuid
document*string

The document ID.

Formatuuid
tenant*string

The tenant identifier

Query Parameters

PageSize?integer
Formatint32
PageToken.Skip?integer
Formatint32
PageToken.Cursor?string

Opaque cursor token from an external system (e.g. RAG API). When set, M:Fruxon.Common.Collections.PageToken.AsString returns this value directly instead of encoding Fruxon.Common.Collections.PageToken.Skip.

Response Body

curl -X GET "https://api.fruxon.com/v1/tenants/string/assets/497f6eca-6276-4993-bfeb-53cbbbba6f08/documents/497f6eca-6276-4993-bfeb-53cbbbba6f08/chunks"
{
  "documentId": "00000000-0000-0000-0000-000000000000",
  "chunks": [
    {
      "chunkId": "00000000-0000-0000-0000-000000000000",
      "chunkIndex": 0,
      "content": "string",
      "contextText": "string",
      "headingPath": "string",
      "pageNumber": "string",
      "startCharIdx": 0,
      "endCharIdx": 0,
      "startLine": 0,
      "endLine": 0,
      "tokenCount": 0,
      "codeIntel": {}
    }
  ],
  "totalChunks": 0,
  "nextPageToken": "string"
}
Empty
Empty
{
  "type": "string",
  "title": "string",
  "status": 0,
  "detail": "string",
  "instance": "string",
  "property1": null,
  "property2": null
}

Lists documents ingested into an asset.

Returns the documents currently materialized in the asset's RAG index, with per-document ingestion status, size, and source metadata. Each document corresponds to a single file or URL pulled from the configured source and is the unit of refresh and deletion. Requires a vectorized asset that has completed at least one ingestion; returns 404 otherwise. Pair with `GET /assets/{asset}/documents/{document}/chunks` to inspect the indexed text.

Deletes a document and its chunks from the knowledge base.

Removes the document and every chunk produced from it from the asset's vector and full-text indices. The next search against the asset will no longer surface this content. The original source is untouched, so a subsequent `POST /assets/{asset}:refresh` (without a `documentIds` filter) will re-ingest the document. Idempotent at the API surface: deleting an already-removed document returns 404.