feat(landing): the console copy says it makes changes

[jfrench9]

Summary

robosystems.ai described a console that only answers questions. The console now also makes changes through /do, so the copy says so. It names no vendor and no model. Changes are described as running on "a stronger model than questions", and connecting your own MCP client stays the lead elsewhere on the page.

Changes

• src/app/(landing)/platform/content.tsx, AI-Powered Interactive Console section:
  • The intro adds that the console makes changes on your own graphs. The "direct queries are always free" line moves here.
  • The second card, Real-Time Execution, becomes Changes with /do. It covers what /do builds, drafts and sorts, that changes run on a stronger model than questions, and that nothing it drafts posts without a person.
• src/components/landing/ApplicationsSection.tsx: the RoboLedger card's console bullet now reads "ask about your books, or have it draft and sort the routine work for you to review".

Breaking Changes

None.

Testing

• Pre-commit gate (709 tests, format, lint, typecheck)
• Not looked at in a browser; copy only.

Companion PRs: roboledger-app #475 (console spotlight, Plaid tile) and a robosystems docs PR.

🤖 Generated with Claude Code

[github-actions]

Change classification: normal — platform or application code changed.

A normal change carries a documented automated review on this pull request before merge, in place of a second human approver. The change-classification job requests that review and passes once it has been posted. Recorded by the change-classification job.

[jfrench9]

@claude please review this PR

[claude]

Claude finished @jfrench9's task in 12s —— View job

---

Review of PR #501

Copy-only change in 2 files. I read the diff. I did not run the tests or look at it in a browser.

Verdict: looks good to merge. I have a few optional notes.

What I checked

• JSX is valid. ' and — are escaped correctly, so the react/no-unescaped-entities lint rule won't fire.
• The edits are text only. There are no new imports, props or logic, so the risk is minimal.
• The copy is consistent across both files. Both say changes need review, and the vendor and model are unnamed as intended.

Notes (non-blocking)

1. Lost "Real-Time Execution" content. The old card described live progress, streaming results, and per-query timing, row counts and credit usage. That is now gone from the page. If those features still exist, you may want them somewhere else, such as the first card or the docs. If they were only filler, this is fine.
2. "Always free" wording. In the intro, "direct Cypher yourself, which is always free" is accurate. The sentence that follows ("On your own graphs it can make changes too…") reads clearly after it, so there's no ambiguity about what is free. Making the free-versus-paid line explicit would help only if /do consumes credits. The "stronger model" phrasing hints at that, and the copy doesn't say whether it does.
3. Scope of the review promise. The card says "nothing it drafts posts without a person". The RoboLedger bullet says "for you to review". Both match, but they are product claims. Please confirm the console enforces this in the product. That includes any API or MCP path, not only the UI.
4. Card length. The new card body is much longer than its siblings. It could look uneven in the three-card grid on desktop. Worth a quick visual check, since the test checklist notes this wasn't viewed in a browser.
   · Branch: feature/console-copy