Returning Pinned Data

What Pinned Data Is For

Use pinned data when you want important outputs to be visible in the observer panel long after the execution completes — for example: the last API response body from an HTTP node, the AI agent's final answer, an extracted data set, or a form submission snapshot. Pinned data is stored in a persistent store (not just in-memory execution state).

Returning Pinned Data

NodeExecutionResult has an OutputBranches field that accepts an INodeResultBranches implementation. The built-in NodeResultBranches class carries both main output and pinned output:

return new NodeExecutionResult
{
    IsSuccess     = true,
    OutputPortKey = "main",
    OutputData    = new Dictionary<string, object>
    {
        { "statusCode", 200    },
        { "body",       body   }
    },

    // Pinned data — persists beyond the execution lifetime
    OutputBranches = new NodeResultBranches
    {
        PinnedData = new Dictionary<string, object>
        {
            { "lastResponseBody",   body            },
            { "lastStatusCode",     200             },
            { "lastRequestAt",      DateTime.UtcNow }
        }
    }
};

Pinned vs Output Data

Null Pinned Data

If you do not set OutputBranches (or set PinnedData to null/empty), no pinned data is written. This is the default for most nodes. Only write pinned data when it adds meaningful value for operators or developers inspecting the workflow.

// Most executors — no pinned data needed
return new NodeExecutionResult
{
    IsSuccess     = true,
    OutputPortKey = "main",
    OutputData    = new Dictionary<string, object> { { "ok", true } }
    // OutputBranches = null (omitted) — no pinned data
};