---
title: "Ingest Document Brain"
method: POST
path: "/api/businesses/{business_id}/documents/{document_id}/brain/ingest"
tags: ["Businesses"]
---

# Ingest Document Brain

`POST /api/businesses/{business_id}/documents/{document_id}/brain/ingest`

## Request

- Base URL: `https://api.flagpost.ai` (the document declares no server; this is the origin it was published from)
- URL: `https://api.flagpost.ai/api/businesses/{business_id}/documents/{document_id}/brain/ingest`
- Auth: HTTP bearer

Queue markdown extraction for one document (internal Brain tab only).

``force`` discards the old read here — pages, chunks, and tables — before
the job exists, so a retried job never wipes what an earlier attempt saved. The job itself only fills
missing pages. The file is out of AI search until the chain is indexed
again: the status leaves ``indexed`` for the whole re-read.

## Path parameters

- `business_id` string, required
- `document_id` string, required

## Query parameters

- `force` boolean

## Response `200`

Successful Response

- object

## Other responses

- `422`: Validation Error

---

[API](https://skmtc.dev/flagpost/apis/flagpost-api.md) - [All operations](https://skmtc.dev/flagpost/apis/flagpost-api/llms.txt) - [OpenAPI document](https://skmtc.dev/flagpost/apis/flagpost-api/revisions/0308cef485ca?raw)