# Make backlog proof: scheduled run and zero-write replay

Silent screen recording with descriptive captions; no speech or audio was recorded. The source capture was made on 18 September 2026. It shows Make results from 17 September 2026.

- 00:00:00.000 — We're checking the same 12 fictional orders in Make.
- 00:00:09.500 — Six should update. Four should stay held and two are already finished.
- 00:00:19.000 — On 17 September 2026, Make started this batch from its schedule.
- 00:00:25.000 — The batch ID is `make-demo-20260917-164705941`.
- 00:00:31.000 — Make updated six rows and held four with reasons.
- 00:00:37.000 — The two processed rows stayed unchanged.
- 00:00:43.000 — Lovable now displays Make's saved batch report.
- 00:00:49.000 — We can see 12 completed rows and four exceptions after a reload.
- 00:00:55.000 — We ran it again. All 12 rows were skipped.
- 00:01:01.000 — Make wrote no input rows.
- 00:01:07.000 — Lovable now displays the saved replay report.
- 00:01:13.000 — It shows 12 skips and no failures.
- 00:01:19.000 — These two runs used 180 Make credits on Core.
- 00:01:25.000 — A separate extra replay used another 80 credits. It isn't part of the 180.
- 00:01:31.000 — Check the empty run before you price the schedule.

The clip is a retrospective walkthrough of two completed fictional runs. Hard cuts remove waiting and navigation. It does not show either run happening live.

The saved-row scene uses the sanitized readback preserved after the baseline. The live Sheet later changed during a separate fallback test, so it is not used here. The native history and Lovable Viewer bind the saved evidence to the same baseline and replay.

The selected 180-credit total is the 100-credit baseline plus the first 80-credit replay. A separate additional 80-credit replay remains visible and is excluded from that total.

This clip proves this fixture and these dated runs. It does not establish ongoing reliability, comparative ease or future credit pricing.