Could we get the API to return mirror nodes in responses in a more useful way? Currently, I believe they are returned with an empty name field, and their own id, with no obvious way to reference the original. This makes the API (and any MCP server built on it) far less useful, since any node could be a mirror, and children nodes that are mirrors are simply empty.Here’s an example of retrieving two nodes, one regular, the other a mirror:
$ curl -s -G https://workflowy.com/api/v1/nodes -H "Authorization: Bearer <xxx>" -d "parent_id=764c7ba1-7432-e6dd-5c53-2ee5dfa7e019" | jq .
{
"nodes": [
{
"id": "154465d1-a3d4-69cc-b3af-5df252cfa584",
"name": "", # <------- EMPTY NAME
"note": null,
"parent_id": null,
"priority": 200,
"completed": false,
"data": {
"layoutMode": "bullets"
},
"createdAt": 1767737682,
"modifiedAt": 1767737682,
"completedAt": null
},
{
"id": "356c1033-e2ed-893b-1c1b-17f21ba0a635",
"name": "A Regular Child",
"note": null,
"parent_id": null,
"priority": 100,
"completed": false,
"data": {
"layoutMode": "bullets"
},
"createdAt": 1767664256,
"modifiedAt": 1767737672,
"completedAt": null
}
]
}
In addition to making the Retrieve and List calls actually accurate, it would allow some analysis on the use of mirror nodes. For example, I sometimes use mirrors when I’m not entirely sure of the ideal location of a node in my graph, indicating perhaps an ambiguity in organization on my part. So, the CLI I wrote could easily produce a list of all mirror nodes, sorted by the reference counts in descending order. See the other reports I have written using my CLI and the API.
Thank you for your consideration.