25 Claude Code SEO Prompts (Copy and Paste)

Fetch https://[your-site.com]/robots.txt and /sitemap.xml. Check that robots.txt doesn't block important pages, CSS, or JS, that it points to the sitemap, and that every sitemap URL returns a 200 status. List problems with the exact line or URL.

For every URL in https://[your-site.com]/sitemap.xml, fetch the page and record the meta robots tag, the canonical URL, and the HTTP status. Flag any page that is noindexed, canonicalized to a different URL, redirected, or missing a canonical. Output a table.

Check every URL in [urls.txt]. Follow redirects and list any URL that takes more than one hop or ends on a 404. For each chain, show every hop and suggest the single final destination to link to instead.

Fetch the raw HTML of https://[your-site.com/page] without running JavaScript. Tell me whether the main heading, body text, internal links, and structured data are present in the raw HTML or only appear after JavaScript runs. Explain the risk for each missing element.

Use the PageSpeed Insights API with my key in $PSI_KEY to test these URLs on mobile: [urls]. Report LCP, INP, and CLS from field data where available, otherwise lab data, and list the top three fixes for each page in plain language.

List every page in this repo with its title tag and meta description. Flag titles over 60 characters, descriptions outside 120–160 characters, duplicates, and titles that don't match the page's H1 topic. Show the file and line for each.

Check the heading structure of every page in [content folder]. Flag pages with no H1, more than one H1, or skipped levels (an H2 followed by an H4). Suggest a fixed outline for the worst five.

Find every image in [src or content folder] with missing, empty, or filename-style alt text (like "IMG_2041"). Suggest descriptive alt text based on the surrounding content, and apply it after I approve.

Here is my Search Console export with clicks, impressions, CTR, and position by page: [file.csv]. Find pages ranking in positions 3–15 with below-average CTR. For each, write two alternative titles under 60 characters that match the top queries and give a reason to click.

Count the words in every post in [content folder] and compare it to its main topic. List posts under 500 words or that don't fully answer the question in their title. For each, recommend expand, merge into another post, or remove.

Create a content brief for "[keyword]". Fetch the top 5 ranking pages: [urls]. Include search intent, the reader's main question, a title under 60 characters, an H2 outline that covers what all top pages cover, internal links to existing pages in this repo, and one angle none of them cover.

Read [content/blog/post.md]. Find statements that may be out of date (old years, deprecated tools, changed prices, dead links). Check each against a current source and propose specific updates with the source URL. Don't change the author's voice.

Compare my page [url or file] with these competing pages: [urls]. List the questions, subtopics, and examples they cover that mine doesn't, ordered by how useful they'd be to the reader. Skip anything that's padding.

Here are "People also ask" questions from Google for [keyword]: [paste questions]. Write a short, direct answer (2–3 sentences) to each that fits my page [file], and add the ones that are genuinely relevant as an FAQ section.

Build a list of every page on this site from the sitemap or the content folder. For each page, count how many other pages link to it. List pages with zero or one internal link, starting with the most important ones.

For [content/blog/new-post.md], find 3–5 existing pages on this site that a reader of this post would genuinely want next. Suggest the exact sentence to add each link to and a descriptive anchor text. Then find 3 older pages that should link to this new post.

List every internal link in [content folder] with its anchor text and target. Flag vague anchors like "click here" or "read more", and flag pages that receive the exact same keyword anchor from many places. Suggest better anchors.

Fetch https://[your-site.com/page] and extract all JSON-LD. List each type, check required and recommended properties against Google's structured data documentation, and flag errors or properties that don't match the visible page content.

Generate JSON-LD for [page or file] as [Article / Product / LocalBusiness / SoftwareApplication]. Use only facts that appear on the page. Mark any property you had to leave out because the page doesn't show it.

Add the JSON-LD above to [file] the way this codebase already handles structured data. If there's no existing pattern, add it as a script tag with type application/ld+json and escape "<" characters. Then build the site and show me the rendered output.

Open [keywords.csv] (keyword, volume, difficulty). Group keywords by search intent and topic so each group could be answered by one page. For each group, suggest a page title, the primary keyword, and whether an existing page in this repo already covers it.

Using the clusters above, build a table mapping each cluster to one URL on this site, or "new page" if none fits. Flag clusters that two or more existing pages are competing for.

Open my Search Console export of queries and pages: [file.csv]. Find queries where two or more of my pages get impressions. For each, show the pages, their clicks and positions, and recommend which page should own the query and what to change on the others.

Fetch https://[your-site.com]/robots.txt. Tell me which of these crawlers are allowed or blocked: GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended, CCBot, Bytespider. Explain what each one does and what blocking it would change.

Rewrite the opening of each H2 section in [file] so it starts with a direct, self-contained answer of 2–3 sentences, followed by the detail. Keep facts and links unchanged. Flag sections where the answer depends on a claim without a source.