NodeFox

Lesson 24 · Advanced

Inspect execution without re-executing it

Understand what recorded playback can and cannot prove.

1:40 video guide + written guide

Source-based instructional visualization · synthetic examples · AI narration. This explains configuration and behavior; it is not a recording of a verified live run.

Jump to a chapter

Before you start

A signed-in NodeFox workspace. Use synthetic data and prepare outputs before enabling external writes.

  1. 01Open a recorded run

    Use the available execution inspection or replay surface for your installed version. The current ReplayView is a read-only projector: its input is a recorded artifact, not a request to run the graph again.

  2. 02Follow the evidence

    Inspect the run outline, graph position and selected firing. Review recorded node configuration, payloads and metadata. An artifact can only show what was recorded; it does not recreate missing evidence.

  3. 03Distinguish playback from rerun

    ReplayView does not start workers, invoke WASM or run network actions. Starting a new execution is a different operation and may call models or repeat external writes. Model outputs can change even when orchestration rules are unchanged.

  4. 04Share carefully

    Review artifacts for sensitive inputs and credentials before sharing. Traceability supports investigation, but it is not a compliance certification or proof that a remote system accepted a proposed action.

If the result is wrong

If an expected field is absent from the artifact, do not infer it from the animation. Inspect the original run or capture configuration.