SEO and content

Restructure a money page answer-first so AI answers can quote it

Pulls the questions the page already gets in Search Console and who AI answers cite, then reorganises the existing content with a direct answer up top, question headings, a comparison table and a page-specific FAQ. Invents nothing, applies after preview and approval, and records a baseline for measurement.

When to use it

Your service or product page ranks in Google but is not quoted in AI answers, or its most important information is buried under a long introduction.

What you get

A new layout built on the page's real questions, with a direct answer up top, question headings and a page-specific FAQ, applied after your approval and verified live, plus a baseline to compare in 4 to 6 weeks.

Does it change anything in my account?

Only with your approval. The assistant first shows a preview of every change and waits for your go-ahead.

Works with

Search Console, SEO Intelligence, WordPress

Fill in before sending

[page URL][Opus Growth hosting or WordPress][Search Console property][price range, duration, warranty and so on, or "none"][domain][date][FACT NEEDED: ...]

Restructure the page [page URL] on my site into an "answer-first" layout that AI answers and Google can quote easily. Site type: [Opus Growth hosting or WordPress]. My Search Console property: [Search Console property]. I think the page's main question is: [main question, e.g. "how much does a boiler service cost in Hackney"]. Facts not on the site that I confirm are correct: [price range, duration, warranty and so on, or "none"]. 1. The questions the page gets. Use get_report with platform='search_console', breakdown='page,query' for the last 90 days and keep only this page's rows. Separate the question-style queries (what, how, why, how much, which, difference, price, and their equivalents in the page's language) with at least 20 impressions, sorted by impressions. Keep these rows as the measurement baseline: query, impressions, clicks, average position. 2. The current content. Read the page with site_read_url, first with outline=true and then with text=true (if truncated is true, read again with max_bytes=1000000). List the heading tree, the first paragraph and every number, price, duration and claim on the page. That list is everything you are allowed to use. 3. What AI says. Check seo_history for earlier runs, then run seo_ai_overview for the main question and the two questions with the most impressions (target=[domain], location and language for my market). Note what the first sentence of the AI answer says, which kinds of facts it contains (price range, duration, steps, comparison) and which pages it cites. If references is empty, count the domains of the links inside ai_answer. These calls cost credits, so do not run more than three without asking me. 4. The new structure. Keep the existing text and propose this layout: 1. The H1 stays, or is lightly adjusted to meet the main question more clearly. 2. Directly under the H1, a 40 to 60 word answer paragraph: a direct answer to the main question in the first sentence, then the two most important concrete facts. No openers such as "Nowadays" or "If you are wondering". 3. The body sections use question-style H2s taken from the real Search Console questions. The first sentence under each H2 answers that question. 4. A table if the page holds comparable facts (packages, durations, options). If it does not, do not invent one. 5. At the end, a 4 to 6 question FAQ specific to this page. The questions come from Search Console or from topics the page really answers. 6. A visible "Last updated: [date]" line and who prepared the page (the business name unless I gave a real person's name). 7. Because this FAQ is visible on the page, FAQPage JSON-LD can be added. Note that since 2023 Google shows FAQ rich results only for well-known government and health sites, so the markup is there for machines to read the content, not for a rich result. 5. No invented facts. Do not write any number, price, duration, warranty, customer count, rating or certificate that is not on the page or in the facts I gave above. If the answer paragraph needs a fact we do not have, write [FACT NEEDED: ...] in its place and ask me. 6. Quality check. Run all new text through seo_content_lint. If human_score is below 80 or any phrase is flagged, rewrite those paragraphs and run it again. 7. Preview. Show the old and new structure side by side, section by section, note which existing fact each new sentence rests on, and wait for my approval. 8. Apply after approval. 1. Hosting: read the page file with site_read_file and change it section by section with site_edit_file. Write the FAQPage block statically in the page head and link it to the site's existing @id graph. 2. WordPress: read the content with wordpress_get_post. If the page is built with Elementor, read the layout with wordpress_page_design and change it with wordpress_edit_page_design; otherwise use wordpress_update_post. Do not change the page's publish status. 9. Verify. With site_read_url and outline=true, show the single H1, the new H2s and the answer paragraph live. With text=true, confirm the answer paragraph is visible without JavaScript. End with a measurement plan: in 4 to 6 weeks we will compare the Search Console rows for the same queries and the seo_ai_overview results for the same questions with the baseline. Do not judge the effect of the change before then.

1. The questions the page gets. Use get_report with platform='search_console', breakdown='page,query' for the last 90 days and keep only this page's rows. Separate the question-style queries (what, how, why, how much, which, difference, price, and their equivalents in the page's language) with at least 20 impressions, sorted by impressions. Keep these rows as the measurement baseline: query, impressions, clicks, average position.
2. The current content. Read the page with site_read_url, first with outline=true and then with text=true (if truncated is true, read again with max_bytes=1000000). List the heading tree, the first paragraph and every number, price, duration and claim on the page. That list is everything you are allowed to use.
3. What AI says. Check seo_history for earlier runs, then run seo_ai_overview for the main question and the two questions with the most impressions (target=[domain], location and language for my market). Note what the first sentence of the AI answer says, which kinds of facts it contains (price range, duration, steps, comparison) and which pages it cites. If references is empty, count the domains of the links inside ai_answer. These calls cost credits, so do not run more than three without asking me.
4. The new structure. Keep the existing text and propose this layout:
   1. The H1 stays, or is lightly adjusted to meet the main question more clearly.
   2. Directly under the H1, a 40 to 60 word answer paragraph: a direct answer to the main question in the first sentence, then the two most important concrete facts. No openers such as "Nowadays" or "If you are wondering".
   3. The body sections use question-style H2s taken from the real Search Console questions. The first sentence under each H2 answers that question.
   4. A table if the page holds comparable facts (packages, durations, options). If it does not, do not invent one.
   5. At the end, a 4 to 6 question FAQ specific to this page. The questions come from Search Console or from topics the page really answers.
   6. A visible "Last updated: [date]" line and who prepared the page (the business name unless I gave a real person's name).
   7. Because this FAQ is visible on the page, FAQPage JSON-LD can be added. Note that since 2023 Google shows FAQ rich results only for well-known government and health sites, so the markup is there for machines to read the content, not for a rich result.
5. No invented facts. Do not write any number, price, duration, warranty, customer count, rating or certificate that is not on the page or in the facts I gave above. If the answer paragraph needs a fact we do not have, write [FACT NEEDED: ...] in its place and ask me.
6. Quality check. Run all new text through seo_content_lint. If human_score is below 80 or any phrase is flagged, rewrite those paragraphs and run it again.
7. Preview. Show the old and new structure side by side, section by section, note which existing fact each new sentence rests on, and wait for my approval.
8. Apply after approval.
   1. Hosting: read the page file with site_read_file and change it section by section with site_edit_file. Write the FAQPage block statically in the page head and link it to the site's existing @id graph.
   2. WordPress: read the content with wordpress_get_post. If the page is built with Elementor, read the layout with wordpress_page_design and change it with wordpress_edit_page_design; otherwise use wordpress_update_post. Do not change the page's publish status.
9. Verify. With site_read_url and outline=true, show the single H1, the new H2s and the answer paragraph live. With text=true, confirm the answer paragraph is visible without JavaScript.

End with a measurement plan: in 4 to 6 weeks we will compare the Search Console rows for the same queries and the seo_ai_overview results for the same questions with the baseline. Do not judge the effect of the change before then.

Open in Claude Open in ChatGPT Discuss in the forum

Example result

An example of the output this prompt produces. Its structure comes from the tools' real output; numbers and names are fictional.

Example output (based on a fictional site, not real customer data)

Baseline (last 90 days, /services/boiler-service/)

Query	Impressions	Clicks	Avg. position
how much is a boiler service	1,840	22	7.4
how often should a boiler be serviced	960	31	5.1
what happens in a boiler service	610	9	11.8

For "how much is a boiler service", Google's AI answer gives a price range in its first sentence and cites two price guide sites. On your page the price sits in the sixth paragraph and only as a "from" price.

Proposed structure
1. H1: Boiler Service in Hackney (unchanged)
2. Answer paragraph (52 words): "A standard boiler service in Hackney starts at £85 and takes about 45 minutes. We clean the burner and heat exchanger and test for gas leaks and flue draw..." Source: price from paragraph six, duration from the "How we work" section.
3. H2: How often should a boiler be serviced?
4. H2: What happens during a boiler service? (the existing eight-point list is kept)
5. Table: standard and full service, duration and scope (both are on the page)
6. FAQ: 5 questions, all from Search Console
7. "Last updated: 11 October 2026, prepared by Lavender Heating"

[FACT NEEDED: price of the full service] is not on the page; send it if you want it added.

seo_content_lint gave the first draft 74 ("in a comprehensive manner" and "take a look" flagged); 91 after edits. After your approval the change was applied; site_read_url outline showed a single H1 and 4 H2s, and the answer paragraph appears in the text=true view.

Measurement: on 22 November the same three queries and the same three questions will be compared with the baseline.

More prompts

SEO and contentChanges after your approvalLive demo

Publish my hosted site in Turkish and English with hreflang, per-language sitemaps and llms.txt

Adds an English version of a static site under /en/: per-language URLs, hreflang and x-default, og:locale and schema language per language, a static language switcher, per-language sitemaps and llms.txt. Checks translation pairing and quality with a script and measures every page on two devices.

When to use itWhen you want to add an English version to your Turkish site hosted on Opus Growth and have each language show up in the right place in search without competing with the other.

My site [site address] on my Opus Growth server is Turkish only for now. I want you to add an English version and set both languages up so that search engines and AI assistants pair them correctly. The audience for the English version: [audience, for example foreign tourists]. Names that are never translated: [brand name, product names, place names]. The language a visitor from elsewhere sees by default (x-default): [Turkish or English].

1. Map the current state. Read site_seo_guide, list the pages with site_list_files and note the restore version with site_versions. Propose an English address for every Turkish page and show the mapping table: Turkish stays at the root (/, /hizmetler/), English goes under /en/ (/en/, /en/services/). Write the addresses in English words, lower case, as directories ending in a slash. Do not change the existing Turkish addresses; if one has to change, 301 it with site_redirects. Wait for my approval.
2. Translation. Write every page in English that reads naturally to the audience, translating meaning rather than sentence by sentence. Keep data such as phone numbers, addresses, prices and hours exactly the same and add no new claims. Titles of 50 to 60 characters, descriptions of 140 to 160, unique in each language. Both languages have the same number of sections. Every English page has `<html lang="en">` and answers the question directly in its first paragraph.
3. Configuration. Add a `locales.en` block to assets/site.config.js: prefix "/en/", hreflang "en", ogLocale "en_US", the English business description, the English menu, call to action, footer note and interface text (skip to content, menu, pages, contact). In an `i18n` block, write the root language, the x-default language, the Turkish and English address pairs, the words that are never translated (keepWords) and a term glossary.
4. Write an i18n-stamp.mjs script and put it in the site root. Its jobs:
   1. Print absolute `<link rel="alternate" hreflang="tr-TR">`, `hreflang="en"` and `hreflang="x-default"` tags on every page between `` and ``.
   2. Inside the block printed by seo-stamp.mjs, set og:locale to en_US on English pages, add the other language as og:locale:alternate in both languages, set the WebSite node's inLanguage to ["tr-TR", "en"] and replace the business node's description with the English one on English pages.
   3. Produce sitemap-tr.xml and sitemap-en.xml with xhtml:link alternates for every address and the file's real modification date as lastmod, and turn sitemap.xml into a sitemapindex pointing to both.
   4. Check the translation and exit with a non-zero code on errors: a page without a pair, html lang not matching the folder, a different phone number or long number in the two languages; as warnings, a different number of h2, h3, images or FAQ items in the two languages, Turkish-character words in the English text outside keepWords, and glossary terms not used.
   Each page's own WebPage node carries that page's language as inLanguage. seo-check.mjs only exempts the root home page from the BreadcrumbList requirement, so add a one-item BreadcrumbList to the /en/ home page.
5. Language switcher. The chrome-stamp.mjs script that prints the menu and footer as static HTML gives pages under /en/ the locales.en menu and interface text, and builds the language switcher from the page's hreflang tags, so every page links to ITS OWN counterpart in the other language, not to the home page. Every language link carries hreflang and lang attributes and a tap target of at least 44 pixels. Do not pick the language by redirecting on the browser language.
6. llms.txt. Write the Turkish /llms.txt at the root and the English version at /en/llms.txt; each gives a short, accurate summary of the site, the list of pages in its own language and the contact details, and links to the other. robots.txt stays open to AI crawlers and points only to the sitemap.xml index.
7. Run order after every edit: `node seo-stamp.mjs`, `node i18n-stamp.mjs`, `node chrome-stamp.mjs`, the Tailwind build, then `node seo-check.mjs`, `node design-check.mjs`, `npx --yes html-validate "**/*.html"`. seo-stamp rewrites its own block every time it runs, so i18n-stamp must ALWAYS run after it. If the site still runs on the Tailwind Play CDN and third-party fonts, move it to compiled CSS and self-hosted fonts first.
8. Measure. core_web_vitals on mobile and desktop for every page in both languages; seo_audit for both versions of the home page. Open the Turkish and English home pages with site_preview at 360, 390, 768 and 1440 pixels wide; English headings are often longer, so look hard for overflow. Criteria: mobile performance at least 95 and the other three scores 100; zero errors from seo-check, i18n-stamp and html-validate.
9. After publishing. Confirm the hreflang tags and the language switcher on the /en/ home page with site_read_url. If my Search Console property is [Search Console property], submit sitemap.xml with gsc_submit_sitemap and check the /en/ home page with gsc_url_inspect.

Output: the Turkish and English address mapping table, a score table per page and language, the i18n-stamp check output, the remaining translation warnings and a list of sentences I should review.

SEO and contentRead-only

Check that AI crawlers see the right language version of your multilingual site

Checks hreflang, canonical and html lang alignment on every language version, the alternates in the sitemap and any automatic redirects based on the visitor's country or browser language. Measures which version Google picks and which language page AI answers cite, without changing anything.

When to use itYour site is published in several languages and one language version does not show up in its own market in Google or AI answers, while a page in another language appears instead.

My site [site URL] is published in these languages: [languages and URL structure, e.g. en: /, de: /de/, fr: /fr/]. My Search Console property: [Search Console property]. The two most important pages per language: [homepage and one service page URL in each language]. One commercial question per market: [one question per language]. This is read-only.

Most AI crawlers and Google's crawlers come from servers in the United States, and their browser language is usually English. If the site redirects visitors automatically by IP country or browser language, these crawlers may never see the German or French page. The point of this job is to measure that.

1. Tag alignment. Read the two pages in each language with site_read_url (default max_bytes; if the body is cut before `</head>`, use max_bytes=1000000). For each page extract the `<html lang>` value, the canonical URL, every `rel="alternate" hreflang` link and x-default. Check: does the canonical point to the page itself (if a language version's canonical points to another language, that version drops out of the index); are the hreflang links reciprocal (if A points to B, B must point to A); does every hreflang URL return 200 (site_read_url with max_bytes=0); are the language codes valid (e.g. "en", "de", "en-GB"; do not mix "de-DE" and "de"); do html lang and hreflang name the same language.
2. Sitemap. Read sitemap.xml and any child sitemaps with site_read_url. Compare the xhtml:link alternates with the on-page hreflang for two pages.
3. Redirect behaviour. Read each language's homepage with site_read_url and max_bytes=0 and note the redirects list. This tool sends requests as a browser from one fixed server, so it does not show what a visitor from another country sees. For that, also measure each language's homepage with seo_technical_audit, which renders the page with JavaScript from servers in another location. If the two tools show a title or H1 in different languages, the site is redirecting visitors from that location to another language version. Flag this at the highest severity and ask me: does the redirect rule exempt search and AI crawlers (Googlebot, Bingbot, GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-User, PerplexityBot), and does it use 302 rather than 301. You cannot see the code or server settings, so do not guess the answer.
4. Google's choice. Inspect the two pages per language in one gsc_inspect_urls call. If Google's chosen canonical differs from the page itself (for example the English page is chosen for the German page), say so. Use get_report with platform='search_console' and breakdown='country' for the last 90 days, then breakdown='page by impressions', and add up impressions and clicks per language folder. Flag any language version that gets almost no impressions in its own market.
5. The language AI sees. Run each market's commercial question through seo_ai_overview with that market's location and language values and target=[domain]. If one of my pages is cited, note which language version it is. If references is empty, use the links inside ai_answer. My English page being cited for a German question is a sign the German page is not being seen enough.

Format the output like this. First a four-sentence summary. Then a page table: URL, html lang, canonical (self or not), hreflang count and reciprocity, x-default, Google's chosen canonical, status. Then a redirect table: language homepage, title in site_read_url, title in seo_technical_audit, difference. Then a market table: language, impressions and clicks in its main market, version cited in the AI answer. End with a fix list in order of severity. If there is an automatic language redirect, add this general recommendation: exempting crawlers from the redirect, or showing visitors a banner that suggests a language instead of redirecting them, is the approach Google recommends for multilingual sites.

SEO and contentRead-only

Build intent-based keyword clusters from a few seed keywords

A ranked list of keyword clusters with volume, difficulty and the right page type or ad group for each.

When to use itWhen you are starting keyword research for a business and only have a few seed keywords.

I sell [product or service] in [city or country]. Start from these seed keywords: [3 to 5 seed keywords].

First check seo_history so you reuse earlier results instead of paying for the same query twice. Then use Keyword Planner and SEO keyword research to collect ideas with average monthly searches, competition and keyword difficulty.

Clean the list:
- remove keywords under 30 monthly searches
- remove job, free, DIY, course and competitor-brand terms unless they clearly fit my business
- merge close variants and plurals

Cluster the rest by search intent, with one cluster per distinct need a searcher has. For each cluster give:
- cluster name and head keyword
- 5 to 10 supporting keywords
- combined monthly volume and average difficulty
- intent: informational, commercial or transactional
- best use: service page, blog post or a Google Ads ad group

Rank the clusters by volume and fit for a small business with a limited budget. Mark the top 3 to start with and explain why in one line each.

SEO and contentChanges after your approval

Turn search terms that convert in ads into organic content clusters

Uses 90 days of converting Google Ads and Microsoft Ads search terms to find topics where you pay for conversions but have weak or no organic visibility. Returns a prioritized content plan and a preview-only list of keyword additions.

When to use itWhen you pay for conversions on search terms but are unsure which topics deserve organic content so you can rely less on ads.

I want to use my ad data to decide what to write about. Use the last 90 days on every source.

1. Pull my Google Ads search terms for the last 90 days (days=90, or an explicit start and end date), sorted by conversions. Keep only terms with at least 1 conversion. If the response is truncated, page through until you have every converting term.
2. If Microsoft Ads is connected, find the default account and pull its search terms for the same 90 days. Keep only terms with at least 1 conversion and add them to the pool, marking which platform each term came from.
3. Group all converting terms into topic clusters by search intent. For each cluster show total cost, conversions and cost per conversion, split by platform where both are present.
4. Find my Search Console property (use [site property] if I named one) and pull organic data for the last 90 days with dimension query, plus a page,query pull so you can see which page ranks for each query. Use a high enough limit to cover the long tail.
5. Match each cluster to the Search Console queries that express the same intent. Use your judgement for close variants and wording differences, and say which queries you counted. Report organic clicks, impressions and average position per cluster, and the page that ranks best for it if there is one.

Classify each cluster:
- Content gap: converts in ads, organic average position worse than 10 or no organic impressions. I should build a page.
- Improve: converts in ads, organic position 4 to 10. Name the existing page that needs strengthening.
- Covered: organic position 1 to 3. Note it as a possible place to test lower ad spend later.

Deliver a table sorted by ad conversions, then the top 5 content-gap clusters with a suggested page title and URL slug for each.

Separately, list my Google Ads keywords (list_entities, entity keywords) and show converting Google search terms that are not yet keywords in the account. For Microsoft terms, just flag them; do not check or change Microsoft keywords. Do not add anything. If you think some Google terms should be added as exact match, tell me which ad group each one belongs in and run add_keywords with confirm=false so I see the exact preview. Wait for my approval before applying.