document_get

Read one file of a brand. An image comes back as an image that the agent can look at.

Ask about one file of a brand. An image comes back as an image that the agent can look at.

“Does this headline work with our logo?”

The agent calls document_get for the logo, and it looks at the image. You then get an answer about the headline and the logo together.

The job guide Describe a brand shows each step of the job.

Reference

Fetch one brand document. Images come back as an image you can actually look at, so you can judge whether copy or creative suits the brand. Everything else — PDFs, oversized images — comes back as metadata and a URL to fetch.

Input

Argument	Type	Required	Description
`document_id`	string	yes	The id of the document to fetch.

Output

A successful call returns this object in structuredContent.

Field	Type	Always present	Description
`document`	object	yes	A finalized document.
`document.createdAt`	number	yes	Milliseconds since the Unix epoch, in UTC.
`document.filename`	string	yes	The name of the file at upload.
`document.id`	string	yes	The id of the document (`doc_…`).
`document.mimeType`	string	yes	The media type of the stored file. AdCrunch reads it from the file at finalize.
`document.name`	string	yes	The display name. It is the filename until someone renames it.
`document.sizeBytes`	number	yes	The size of the stored file, in bytes. AdCrunch reads it from the file at finalize.
`document.targetId`	string	yes	The id of the brand that holds the document (`brd_…`).
`document.targetType`	string	yes	The kind of record that holds the document: `brand`.
`document.url`	string	yes	The public URL of the file. Each person who has the URL can read the file.

Failure codes

A failed call has isError set, and structuredContent.error holds one of these codes. Errors describes the shape of a failed call.

not_found
forbidden
invalid_request
internal_error

Scope

The token must hold brand:read. Auth & scopes lists each scope.

Annotations

A client reads these hints. A hint that the tool does not declare has the default value of the MCP specification.

Read-only. The tool changes nothing.
Closed world. The tool reads and writes the data of AdCrunch only.

Example

The arguments:

{
  "document_id": "doc_7d1f58"
}

The result, in structuredContent:

{
  "document": {
    "createdAt": 1790086400000,
    "filename": "acme-logo.png",
    "id": "doc_7d1f58",
    "mimeType": "image/png",
    "name": "acme-logo.png",
    "sizeBytes": 48213,
    "targetId": "brd_6a9c33",
    "targetType": "brand",
    "url": "https://documents.adcrunch.dev/org_3a7f10/doc_7d1f58"
  }
}