get-assets

Use get-assets to look up the images or videos on a page, for example to work out which one the user means before changing it. Asset IDs come from the fills returned by start-editing-transaction.

"Which of the photos on slide 3 is the team photo? Swap that one out."

Pass asset_ids, an array of image or video IDs.

IDs, URLs, and tokens below are placeholders; field names and structure are exactly as returned.

The result is a list of content blocks rather than one JSON value. For each requested ID, in the order you supplied them, there's a text block holding that asset's JSON, followed by an image block with the asset itself when the lookup succeeded. A request for two IDs where the second can't be found returns three blocks:

1. A text block for the first ID:

   {
     "asset": {
       "type": "image",
       "id": "ASSET_EXAMPLE_1",
       "name": "Team photo",
       "tags": ["team", "offsite"],
       "created_at": 1758000000,
       "updated_at": 1758000000,
       "owner": { "user_id": "USER_EXAMPLE", "team_id": "TEAM_EXAMPLE" },
       "thumbnail": {
         "width": 595,
         "height": 335,
         "url": "https://document-export.canva.com/EXAMPLE_THUMBNAIL"
       },
       "import_status": { "state": "success" },
       "metadata": { "type": "image", "width": 1920, "height": 1080, "smart_tags": ["people", "outdoors"] }
     }
   }

   JSON

   Copy

2. An image block containing the first asset's image data.

3. A text block for the second ID:

   { "error": { "code": "asset_not_found", "message": "<human-readable description>" } }

   JSON

   Copy

asset.type is image or video, and metadata carries the dimensions (and duration for videos) when available. The call succeeds as a whole even if individual assets fail, so check each text block.

• MCP tools and rate limits
• start-editing-transaction
• perform-editing-operations
• upload-asset-from-url
• Verify your Canva MCP app