Skip to main content
Structured traversal from one node. Use it instead of pulling every node flat, and after a search when a candidate parent needs a closer look.

Parameters

Every node in a response carries id, type, parentId, data, and createdAt.

up: ancestor chain

Leaf to root, starting with the node itself.

down: subtree

Depth-first, the node itself first at depth: 0, then every descendant with its depth.
A subtree over 50 nodes adds a _tip suggesting a narrower direction or a filtered nodes read.

lateral: siblings

Nodes with the same parent, excluding the queried node. A root node has no siblings; the response says so in _meta._tip rather than returning an empty list silently.

context: everything around one node

Two details to read carefully:
  • ancestors includes the node itself as its first entry, then walks up to the root.
  • children is the entire subtree below the node, not just direct children. Each entry has a depth (direct children are depth: 1).

Errors

  • 400 { "error": "direction must be one of: up, down, lateral, context" }
  • 404 { "error": "Node 'job-999' not found", "_meta": { "_tip": "Never guess node IDs. ..." } }
Node IDs are never worth guessing. Find them with Search.