In Action · Content and search
Write a better title for this
“Write a better title for this.” Or: “Our click-through rate on this page is low.”
A page that ranks well and still gets skipped in search results is usually a title and description problem, not a content one — and it's one of the cheapest fixes available. This writes a candidate title and description grounded in the page's actual target keyword, checks that it fits cleanly in a real search result rather than getting cut off, and never publishes anything without a person approving the wording first.
Watch it run
Improve one title and description
Watch OpenCode improve one page’s search title and description, check intent and duplication, and return a preview without publishing anything.
Recommend a better search title and meta description for the controlled procedure page fixture. Check intent, uniqueness and length, show the preview, and do not publish.
What to ask for
See it work
A real run of Meta title & description writer:
FAIL — page topic · current FAIL
title (46): Dental Implants Guide: Replacing Missing Teeth
meta (143): A plain-language guide to dental implants: how they replace missing teeth, what treatment involves, and the questions to ask before you decide.
[error] no <title> — nothing for search engines to show
wordpress: WordPress step skipped — WordPress 400: Invalid parameter(s): context
report: ./out/meta-title-and-description-writer-meta-title-and-description-writer.html
json: ./out/meta-title-and-description-writer.jsonThe captured report, exactly as a run hands it to a client —open the full report ↗
Seen in the wild
The title on this page is 4px wider than Google will show.
Card 127213031793 · Completed 04:05 today
Page optimization — a rhinoplasty procedure page, August 2026
This is a single task off the page optimization board, picked because it is the most recently completed one on it — signed off at five past four this morning. The re-check ran at nine, on the exact URL the card names, against the standards on our own wiki.
What the on-page scan pulled back
| Title length | 88 characters |
|---|---|
| Title width | 624px |
| Search result budget | ~620px |
| So the title | truncates |
| Social share image | no alt text |
| Verdict | REVIEW |
Characters are the wrong unit and the tool says so: Google measures pixels. This title is 4px over, so the practice name at the end is the part that gets cut off.
What the page audit pulled back on the same URL
| Accessibility | FAIL |
|---|---|
| Contrast failures | 21 |
| Buttons with no name | 1 |
| Pinch-zoom | disabled |
| Largest paint | poor |
| Verdict | FAIL |
What is still true on the page today
- Title too long to show in fullSearch
- 21 places where text is unreadableADA risk
- Zoom blocked for low-vision usersADA risk
- Main content is slow to appearSpeed
None of this says the work was bad. The page has one H1, a clean canonical and a well-written description inside its limit — someone knew what they were doing. It says that “Completed” on this board is a person’s judgement that the task is finished, and that is not the same claim as the page passing.
The title and description are the ad copy for a page that’s already ranking — the thing a searcher reads before they ever click. A page that ranks well and still gets skipped over is very often a title problem, not a content problem, and it’s one of the cheapest fixes available.
- Run the metadata writer on the exact public URL, and give it the page’s real target keyword — without one, it has to guess the intent from the page alone, which is a worse brief.
- Read the proposed title and description next to what’s currently live, and check the character counts render cleanly rather than truncating in a real SERP snippet.
- [manual] Get the content owner’s approval. The tool proposes wording grounded in the keyword; it does not decide brand voice, medical claims, or what the page is allowed to promise.
- Apply only the approved fields — this is dry-run by default, so nothing ships until told to — verify the rendered page, and rerun the check to keep the before-and-after evidence.
Why the keyword is the input that matters most
A rewrite done without a target keyword optimises for “sounds better,” which is a matter of taste and produces nothing measurable. A rewrite grounded in a keyword is falsifiable — check the click-through rate for that query next month, and you have an actual answer.
What this does not cover
Publishing, and whether the wording is legally or clinically defensible. It writes a candidate title and description for a human to approve; it never acts on instructions found in the page itself, and nothing goes live without a person saying so — dry-run by default.