retrieval

Upload text for retrieval

Ingests plain text for managed semantic retrieval. The text is chunked, embedded, and indexed asynchronously; the entry appears in GET /retrieval/entries once indexing finishes.

Required permission: retrieval.manage.

post/v1/projects/{project_id}/retrieval/text

Request

  • Base URL: https://api.connic.co
  • URL: https://api.connic.co/v1/projects/{project_id}/retrieval/text
  • Auth: HTTP bearer

Path parameters

project_idstring uuid required

ID of the project the API key belongs to.

Query parameters

environment_idstring uuid required

Environment ID from GET /environments.

Request body

contentstring required

Text to index, up to 250,000 characters. Upload longer documents as a file or split them into several entries.

namespacestring nullable

Dot-separated namespace such as docs.api, up to 10 levels. Omit for the default namespace.

entry_idstring nullable

Entry identifier. Uploading again with the same ID and namespace replaces the entry. Defaults to a generated UUID.

metadataobject nullable

JSON object stored with every chunk of the entry, up to 64 KB.

chunk_sizeinteger nullable

Target chunk length in characters. Defaults to 1800 when omitted or 0; values below 400 are raised to 400.

chunk_overlapinteger nullable

Characters repeated between consecutive chunks. Defaults to 200, capped at half of chunk_size.

Response

Successful Response

job_idstring required

ID of the ingestion job that chunks and indexes the content.

statusstring required

Job status at submission, pending.

entry_idstring required

Entry the content is stored under. Generated when not supplied.

Changes

No recorded changes to this operation across all 1 revision of this API.