Not a restated error. A verdict you can act on, with everything you need to check it yourself.

An example diagnosis

Illustrative example: a mapping references a property absent from the supplied Iterator output. This example is not a verified customer resolution.

What happened

  1. Make reported the failure on module 04, Google Sheets — Add a Row.
  2. The execution trace showed module 03, the Iterator, returning metadata only.
  3. The mapping on the Name field resolved to empty, and the upstream output contained no such property.
  4. Observation: the referenced property is absent. Check the actual output structure and mapping before concluding why.
  5. The change: Name ← iterator.value.filename instead of Name ← iterator.name.

This is an illustrative scenario, not a measured customer result or a guaranteed resolution time.

Evidence trail

Available sources and observations are shown so you can review the reasoning.

A diagnosis is only worth as much as what backs it, so each line points at where it came from: the field, the bundle index, the module number, the value that was actually resolved on the failed run.

Why this is not decoration

  • It lets you disagree. If the evidence does not support the conclusion, you can see that in seconds.
  • It shows the difference between reading and guessing — an inference is labelled as one.
  • It is what makes a confirmed verdict mean something rather than sound confident.
Rule

If a claim cannot point at data you provided, it is stated as an inference, never as a fact.

History

Saved diagnoses can be searched within the history window of your plan.

Search by error, by module, or by scenario. Each entry keeps the root cause, the certainty level, and when it was resolved — so a failure you saw two months ago is a lookup, not a new investigation.

Review past analyses

Search and review previous analyses manually. Automatic cross-scenario pattern detection is not currently available.

Inconclusive?

Sometimes several causes remain compatible with what you provided. FixScenario says so.

It lists what it ruled out and why, keeps the remaining candidates ranked by plausibility, and names the one piece of missing information that would settle it — the execution log for a given run, the full HTTP response body, whether a connection is shared with another scenario.

How credits work

Each completed analysis uses one of your 3 free credits, including an inconclusive result. On a paid plan, inconclusive runs do not count against your diagnostic quota. Guessing would cost you more than silence: you would go and change a field that was never the problem, and then have two things to undo instead of one.

Why this matters

A tool that always sounds certain is a tool you cannot calibrate. The value of confirmed depends entirely on the existence of inconclusive.

Something broke right now?

Paste the error. Your first 3 completed analyses are free.

Diagnose my workflow