Dropbox MCP Server

Dropbox

Dropbox is cloud file storage with a single path namespace per connected account. Paths look like /Reports/Q1.pdf; the account root is /.

First moves

1. Current account — confirm which Dropbox user is connected.
2. Space usage — check allocated and used storage before large downloads or bulk uploads.
3. List a folder — browse / or a known folder path; paginate when more entries remain.
4. Search — find files by name or query when the folder path is unknown.
5. Metadata — read size, revision id, and content hash before downloading large files.
6. File content — pull inline text for documents; binary files return encoded content or a download link.
7. Download — fetch raw file bytes when the workflow needs the full binary payload.
8. Shared links — audit existing share urls on a path before creating new ones.
9. Revisions — review file history before rolling back an older version.

Data model

Account root → folders → files. Shared folders appear under the same paths the connected user sees in Dropbox. Each file has a revision id that moves forward on every save.

Operational order

1. List or search to discover paths — do not guess file locations from memory.
2. Read metadata before downloading large files or comparing versions.
3. List shared links on a path before creating a new public or team link.
4. List revisions before restore — pick the revision timestamp the user describes.
5. For organizational moves, confirm source and destination paths exist.

Writes

Create folders, upload or overwrite files, move and copy paths, delete files and folders, create and revoke shared links, and restore an earlier revision. Restore and delete overwrite or remove content — expect confirm_write on destructive actions.

Gotchas

• Folder pagination — large directories return a cursor; continue listing instead of stopping at the first page.
• Search lag — very new uploads may take a few seconds to appear in search results.
• Large downloads — big binaries may return a signed download link instead of inline content.
• Shared link hygiene — list existing links before create to avoid duplicate urls on the same path.
• Revision restore — rollback replaces the current file; list history first to match the user's date language.
• Path existence — list or search confirms a path before move, copy, or delete.
• Rate limits — paginate deep folder walks instead of fanning out across many shallow lists at once.