# PR5.63 — Owner Chrome checker follow-up

Research only on clean baseline 4673ef9. Current accepted checkpoint: 18 products, 8 routes, 2 remaining. No graph, claim, engine, SQL or publication changes. PR5.62 and original corpus records remain unchanged.

## Returned observations

The owner opened the official public passive checker in Chrome. The header shows My Grammarly and Finish account setup; do not describe this as an anonymous test or assume an exact plan from the header. The fixed passive sentence was visible without a readable passive suggestion.

The separate spelling control, This is a sentnce., produced a red underline and a readable Correct your spelling card proposing sentence, with an Accept button and suggestion counter 1. The right panel still displayed introductory instructions. Therefore introductory panel text alone is not evidence of an inactive checker. Spelling feedback liveness is now established in this owner environment; it was not established in the earlier IAB probe. These are distinct environments, not a retroactive correction of the historical observation. Browser-extension versus embedded-widget feedback was not isolated. The passive screenshot remains a bounded no-visible-feedback observation, not a universal absence.

In the separate signed-in Editor, the same passive sentence produced a readable Improve your text rewrite: The editor wrote the report before sending it to the team. The card explicitly said Pro suggestion · 3 free samples left today. This is observed limited Pro sampling, not ordinary-Free rule entitlement or evidence of a universal allowance/reset schedule. No accepted rewrite or active-after result was returned.

## Decision

Stop this owner probe; do not request another five-pair corpus, spend more samples, change settings or buy a plan. Keep the two writing routes unapproved and zero-qualified. Real Write remains a separate passive-detection research lead with its prior metric anomaly; no general Grammarly-replacement or privacy claim follows.

The next step is an explicit scope/edition decision before further candidate testing: keep the original Free anchor evidence frozen and separate any passive-only comparison from the existing route. Neither spelling success nor a Pro preview can silently resolve the matching-operation gate. No public changes are proposed in this packet.

Five owner screenshots are delivered in evidence/ outside the Git payload. Structured record pins their bytes and attachment names. All conclusions here derive from returned screenshots; no new live vendor research was performed.