Uploadcare MCP. Manage media assets, from metadata checks to bulk deletion.

You manage your whole media stack—uploading, checking metadata, grouping assets, cleaning house—straight through your AI client. You don't have to touch a dashboard; you just tell your agent what to do.

Getting the Raw Data:

To figure out what files you’ve got and how big they are, start with list_files. This tool gives you an inventory of every asset stored in the project and lets you paginate through massive lists using a limit parameter. For specific info on one file ID, run get_file_details; that pulls technical metadata like dimensions, MIME types, and all the raw CDN URLs associated with it.

Need to know about whole collections? You can use list_file_groups first to see every immutable group available in your project, then dive deeper into those specific files using get_group_details.

Handling Assets: Storing and Deleting:

When you’re done with a temporary file or need to promote it to permanent storage, you've got two options. For single assets, use store_file to mark that one temp item as permanent within your account. If you've got a bunch of temps waiting to be solidified, run batch_store_files to mark multiple temporary assets for permanent survival past any standard cleanup cycles.

Conversely, when you need to wipe something out, the tools are brutal about it because they should be. To permanently delete one specific file and all its derived CDN variants, use delete_file. If you’re doing a massive purge of several files at once, don't sweat it; batch_delete_files irrevocably removes multiple specified files and every single variant generated from them in one go.

Moving and Copying Assets:

You can move assets where you need 'em. You tell your agent to use copy_file, which copies an existing file either to a local destination or to a remote cloud target, like S3. This lets you duplicate files across different storage systems without re-uploading.

Monitoring Your Usage:

You gotta keep tabs on your bills and your quota. Run get_project_info when you want real-time data showing how much total storage you've used (in GB) and what your remaining bandwidth consumption stats are. That’s your quick check to see if you're running low or burning through capacity.

The Workflow:

Your agent can handle the whole lifecycle: you list files with list_files, grab technical specs on a specific ID using get_file_details and group contents via get_group_details. If those assets are good, you might use copy_file to send them offsite. Once they're safe, you can either permanently mark 'em using batch_store_files, or if they’ve served their purpose, wipe the slate clean using batch_delete_files for a complete removal of all associated data.