POST /v1/brain/ingest

Ingest documents into the brain (document items only; conversations go through POST /v1/brain/remember).

POST /v1/brain/ingest

Ingest documents into the brain: each item becomes an extractable note plus extract/enrich jobs. Up to 100 items per call. Conversation/session capture goes through POST /v1/brain/remember instead.

Required scope: brain:write

Body fields

Name	Required	Notes
`items`	yes	Array (max 100) of document items.
`items[].type`	yes	Always `"document"`.
`items[].content`	yes	Markdown/text body.
`items[].title`	no	Document title.
`items[].path`	no	Target path; defaults to a content-derived path under `/private/notes/`.
`items[].tags`	no	Frontmatter tags.
`items[].visibility`	no	`"private"` (default) or `"workspace"`.
`items[].sourceRef`	no	Stable id — re-ingesting the same sourceRef is idempotent.
`items[].sourceKind`	no	Integration identity ("mail", "linear", …) — routes the doc to `/private/sources/<sourceKind>/`.

Example

curl -X POST 'https://brain.unisonlabs.ai/v1/brain/ingest' \
  -H "Authorization: Bearer $UNISON_TOKEN" \
  -H 'content-type: application/json' \
  -d '{"items": [{"type": "document", "content": "# Meeting notes...", "title": "Kickoff", "sourceRef": "meeting:2026-07-01"}]}'

Response: { "items": [{ "type": "document", "docId": "...", "path": "...", "jobIds": ["..."] }] } — or an error envelope { "error": { "code", "message" } } on failure.