Resources¶

The MCP resources this server registers are listed below. Collection resources return JSON; document resources take a document ID.

Collection resources¶

URI	Description
`config://paperless`	Server configuration snapshot: the Paperless API URL, the public UI URL and the default page size for list tools
`stats://paperless`	Paperless-NGX document statistics
`remote-version://paperless`	The newest Paperless-NGX release published upstream, and whether it is newer than the connected instance. Not the version installed on it
`tags://paperless`	All tags
`correspondents://paperless`	All correspondents, each with `last_correspondence`
`document-types://paperless`	All document types
`custom-fields://paperless`	All custom fields
`storage-paths://paperless`	All storage paths
`saved-views://paperless`	All saved views
`tasks://paperless`	First page of unacknowledged background tasks

The task resource returns a JSON array. Task timestamps are ISO 8601 strings; missing timestamps are null. Use the list_tasks tool to request further pages or filter tasks.

Document resources¶

URI	Description
`paperless://documents/{document_id}`	Document metadata by ID, without OCR content
`paperless://documents/{document_id}/content`	First 20,000 characters of extracted text. A partial result points to `get_document_content` for paging and to transfer links when available.
`paperless://documents/{document_id}/metadata`	File metadata: original filename, checksums, MIME type
`paperless://documents/{document_id}/notes`	Notes attached to the document
`paperless://documents/{document_id}/history`	Audit history
`paperless://documents/{document_id}/thumbnail`	Thumbnail image

On HTTP deployments with PAPERLESS_MCP_BASE_URL, original files, archived PDF files, and previews are available through create_download_link. Returning whole files as resources puts their bytes in model context, so the inline preview and download resources are not registered.