thinkhard.SI

Prompt repository Marketing

Check a draft article against its target search query before you publish

You have a finished draft and the search query it should rank for, and want an editor's check of intent, gaps, headings and links before it goes live.

Registry no.
MKT-068
Collection
Marketing
Laid out for
Any chat model
Score
10 out of 10
Blanks to fill
6

MarketingAny chat

The prompt

Score 10 of 10, Excellent

Use it when You have a finished draft and the search query it should rank for, and want an editor's check of intent, gaps, headings and links before it goes live.

  • {{query}}
  • {{searcher_need}}
  • {{length_limits}}
  • {{draft}}
  • {{top_results}}
  • {{our_pages}}
Full prompt and why it works
Prompt
Review the draft below against the search query it is meant to rank for, and tell me what to change before we publish. The reader is someone searching "{{query}}" who wants {{searcher_need}}. The review is for me, the editor, so I can make the fixes in one pass.

Title and description length limits we work to: {{length_limits}}

The draft:
"""
{{draft}}
"""

What the top results for this query cover and how they are laid out, from my own search:
"""
{{top_results}}
"""

Our pages that could be linked from this article (title and URL):
"""
{{our_pages}}
"""

Use these headings, in this order:
1. Intent match: does the draft give the searcher what they came for, in the format they expect, such as a how-to, a list, a comparison or a definition? Quote the sentence where the answer first appears and say whether it should come sooner.
2. Answer first: do the first 100 words answer the query directly? If not, write a replacement opening using only the draft's own content.
3. Gaps: questions the top results answer that the draft does not, and points only our draft makes that deserve more space.
4. Headings: whether each H2 says what its section answers, with rewrites where a heading is vague.
5. Internal links: two to five from our list, each with anchor text and the sentence it belongs in.
6. Title tag and meta description: two options each, within the limits, accurate to the article.
7. Trust: claims and numbers that need a source or a date, and advice that needs the author's experience behind it.

End with a table of changes with the columns Change, Where, Why, Effort (small or large).

A good review should change what helps the searcher and leave the writer's voice and strong sections alone.

For example, a heading rewrite in the right form: "Costs" becomes "How much a new roof costs, and what changes the price". That example shows the form only.

Do not repeat the query in every heading, pad the article to match competitors' length, or add facts, statistics or quotes that are not in the draft. If the draft cannot match the intent without a full rewrite, say so in one line at the top instead of patching it.

Variables

{{query}}
The search query the article targets, for example "how to price a cleaning job".
{{searcher_need}}
What the searcher wants, for example "a simple formula and a worked example they can copy".
{{length_limits}}
Your limits, for example "titles under 60 characters, descriptions under 155".
{{draft}}
The full draft, with its headings.
{{top_results}}
For each of the top results: its title and headings, with a line on what it covers.
{{our_pages}}
Your existing pages that could be linked, one per line with title and URL.

Why it works

  • What good looks like. The check starts from the searcher's intent and whether the answer comes early, then works down to headings, internal links, the title tag and claims that need a source.
  • What to avoid. No keyword stuffing, no padding to match competitors' length and no new facts, so the fixes improve the draft without changing what it claims.
  • If unsure. If the draft needs a full rewrite to match the intent, the review says so at the top instead of patching it section by section.

How to use it

  1. Press Fill in blanks and type your details into each blank. What you type stays in this browser.
  2. Copy the filled prompt into the AI tool you already use, such as ChatGPT, Claude or Gemini, or open it in the Grader or the Prompt Tester here.
  3. Before you rely on it, try it on three or four real examples. The Prompt Tester runs them all at once and shows which ones go wrong.
Use it from code

Get this prompt as JSON from the free REST API, with your own text in the blanks, or let an AI assistant fetch it through the MCP server. No key.

curl -s https://thinkhard.si/api/v1/prompts/seo-draft-check-before-publishing

With the MCP server added to Claude, ChatGPT, Cursor or VS Code, call get_prompt with {"prompt": "seo-draft-check-before-publishing"}, or just ask for it by name.