---
title: "Public Extract"
method: POST
path: "/api/ai-fluency/{magic_token}/extract"
tags: ["ai-fluency"]
---

# Public Extract

`POST /api/ai-fluency/{magic_token}/extract`

## Request

- Base URL: `https://api.ruzora.com` (the document declares no server; this is the origin it was published from)
- URL: `https://api.ruzora.com/api/ai-fluency/{magic_token}/extract`
- Auth: none declared

An uploaded session export or patch/zip → text for the form.

Multipart form: ``kind`` (diff | agent_log) and ``file``.

Only for an OPEN task (the token is the credential, as for submit).
Nothing is stored: the page puts the text in the field, the candidate
can read and edit it, and submit stores it exactly as a paste.

The form is parsed BY HAND, after the size and token checks: declaring
``File(...)`` / ``Form(...)`` params makes FastAPI spool the whole body
(to disk past 1 MB, with no cap) before this function runs, so the 2 MB
limit would only ever be checked after an attacker's gigabyte had been
written. Content-Length is required (every browser FormData upload sends
it) and the server never reads past it.

## Path parameters

- `magic_token` string, required

## Response `200`

Successful Response

- unknown

## Other responses

- `422`: Validation Error

---

[API](https://skmtc.dev/ruzora/apis/ruzora-candidate-crm-api.md) · [All operations](https://skmtc.dev/ruzora/apis/ruzora-candidate-crm-api/llms.txt) · [OpenAPI document](https://skmtc.dev/ruzora/apis/ruzora-candidate-crm-api/revisions/f9b8223f3ac4?raw)