Free local n8n diagnostic

Find the data-shape change between manual and production runs.

Paste two redacted JSON samples. The comparison runs entirely in this browser and reports paths and types—not the values inside them.

No uploadNo signupStructure only

Redact before pasting

Compare one equivalent item from each environment.

Remove credentials, tokens, personal data, and customer data. Synthetic replacements preserve structure just as well.

Safety limits: 500,000 characters, 5,000 inspected nodes, and 16 nesting levels per sample.

How to use the evidence

Diff the first node where the two runs diverge.

  1. 01

    Pin comparable data

    Use the same logical input and capture the output of the same node in the manual and production executions.

  2. 02

    Repair the earliest structural change

    Check mapping, expression, node version, preceding branch, and source-system payload before changing downstream nodes.

  3. 03

    Prove the real outcome

    Rerun the active workflow and verify the final CRM, database, message, booking, or other external state—not only a green run.