[DON’T HAVE TIME? READ THIS]
Need ChatGPT prompts for technical SEO audits?
Here are 5 copy-paste prompts that actually work:
- Crawl budget analysis – Identifies pages wasting crawl budget (large files, duplicate URLs, redirect chains)
- Core Web Vitals diagnostic – Pinpoints specific LCP, FID, and CLS issues with fix recommendations
- Schema markup validator – Checks JSON-LD for errors and suggests improvements
- Internal linking audit – Finds orphan pages, broken links, and link equity distribution problems
- Mobile optimization checker – Identifies mobile-specific issues (viewport, tap targets, font sizes)
How to use them: Copy the prompt, replace [YOUR URL] with actual URL, paste into ChatGPT Plus with browser/code interpreter enabled.
What you need: ChatGPT Plus ($20/month) or access to GPT-4. Some prompts work better with MCP integrations like Ahrefs or Technical SEO MCP.
Bottom line: These prompts turn ChatGPT into a technical SEO assistant that spots issues faster than manual audits. Won’t replace proper crawl tools but speeds up initial diagnostics.
I’ve been testing ChatGPT prompts for technical audits over the past few months. Most are too generic to be useful. These 5 actually find real issues.
Crawl Budget Analysis Prompt
Crawl budget matters more than people think. Google wastes time on useless pages, misses important ones.
Copy this prompt:
text
Analyze this URL for crawl budget issues: [YOUR URL]
Check for:
– Pages with large file sizes (over 1MB) that slow crawling
– Duplicate URLs with different parameters
– Redirect chains (3+ redirects to reach final destination)
– Pagination issues causing infinite crawl spaces
– URLs blocked in robots.txt that shouldn’t be
– Faceted navigation creating duplicate content
For each issue found:
1. Explain why it wastes crawl budget
2. Estimate severity (high/medium/low)
3. Give specific fix recommendation
Format as a prioritized action list.
When to use it:
Large sites (10,000+ pages) where Googlebot might miss important content. E-commerce sites with faceted navigation. Sites with complex URL parameters.
I tested this on a client’s e-commerce site with 50,000 products. ChatGPT flagged 15 different filter combinations creating duplicate product pages. Google was crawling 200,000+ URLs for 50,000 actual products.
The fix: robots.txt rules and canonical tags. Crawl efficiency improved within two weeks.
Pro tip: If you have Google Search Console data, paste the “Crawl Stats” report into ChatGPT along with this prompt for more accurate analysis.
Core Web Vitals Diagnostic Prompt
Core Web Vitals affect rankings. But PageSpeed Insights gives you 50 recommendations – which ones actually matter?
Copy this prompt:
text
I need a Core Web Vitals diagnostic for: [YOUR URL]
Analyze these metrics:
– Largest Contentful Paint (LCP) – target under 2.5s
– First Input Delay (FID) – target under 100ms
– Cumulative Layout Shift (CLS) – target under 0.1
– Interaction to Next Paint (INP) – target under 200ms
For each failing metric:
1. Identify the specific element causing the issue
2. Explain why it’s slowing performance
3. Provide 3 actionable fixes ranked by impact
4. Estimate improvement after implementing each fix
Focus on issues that affect actual user experience, not just lab scores.
When to use it:
Pages with slow load times. High bounce rates despite good rankings. After major site redesigns. Before pushing pages for featured snippets (they need good CWV).
I ran this on a client’s product pages that ranked #3-5 but weren’t converting. ChatGPT identified hero images causing LCP issues (4.2 seconds). Recommended lazy loading above-the-fold images was wrong – we actually needed to preload the hero image.
Fixed it. LCP dropped to 1.8 seconds. Conversions up 23%.
Pro tip: Use this with actual PageSpeed Insights data. Paste the full report into ChatGPT along with this prompt. It’ll prioritize fixes that matter most.
Schema Markup Validator Prompt
Schema markup is messy. One wrong property breaks rich results. Google’s Rich Results Test shows errors but doesn’t always explain fixes.
Copy this prompt:
text
Validate this schema markup: [PASTE YOUR JSON-LD CODE]
Check for:
– Required properties missing for the schema type
– Incorrect data types (string vs integer vs boolean)
– Deprecated schema types that should be updated
– Properties that don’t belong to this schema type
– Opportunities to add recommended (not required) properties
Then suggest:
1. Complete corrected version of the JSON-LD
2. Additional schema types that would benefit this page
3. How to nest schema types correctly if multiple are needed
Be specific about which Schema.org version and Google requirements you’re following.
When to use it:
Before publishing new content types. After rich results disappear. When adding complex nested schema (like Recipe with Review and AggregateRating).
I used this for a client’s local business schema. They had LocalBusiness markup but were missing OpeningHoursSpecification and geo coordinates. ChatGPT caught it, provided corrected code.
Rich results appeared within 5 days.
Better version: If you’re using Amanda Jordan’s Schema Advisor custom GPT (mentioned earlier), it’s trained specifically on Schema.org and will give more accurate recommendations.
Internal Linking Audit Prompt
Internal linking structure affects how PageRank flows. Most sites have orphan pages, broken links, and weird link distribution.
Copy this prompt:
text
Analyze internal linking structure for: [YOUR DOMAIN]
Identify:
– Orphan pages (pages with zero internal links pointing to them)
– Pages with broken internal links (404 errors)
– Redirect chains within internal links
– Pages with excessive internal links (over 100)
– Important pages with too few internal links (under 3)
– Link equity distribution – which pages get the most/least links
Provide:
1. List of orphan pages that should be linked from relevant content
2. Suggested anchor text for new internal links
3. Pages that need more internal links to boost their authority
4. Pages that should link to each other based on topic relevance
Consider semantic relevance when suggesting link placements.
When to use it:
Large content sites with hundreds of blog posts. After site migrations. When pages aren’t ranking despite good content.
I tested this on our own blog (about 200 posts). ChatGPT found 12 orphan pages – mostly older posts that got buried. Also found our highest-traffic post only linked to 3 other articles.
Added 8 contextual internal links from that post. Traffic to linked pages increased 40% within a month.
Limitation: ChatGPT can’t actually crawl your site. You need to feed it data from Screaming Frog, Ahrefs Site Audit, or similar tools. Export the internal links report, paste into ChatGPT with this prompt.
If you have Ahrefs MCP set up, you can directly query your Site Audit data through ChatGPT without manual exports.
Mobile Optimization Checker Prompt
Mobile-first indexing means mobile issues = ranking issues. But mobile problems are different from desktop.
Copy this prompt:
text
Check mobile optimization for: [YOUR URL]
Analyze:
– Viewport configuration (is width=device-width set correctly?)
– Touch target sizes (are buttons/links at least 48×48 pixels?)
– Font sizes (readable without zooming?)
– Horizontal scrolling issues
– Interstitials or popups blocking content
– Mobile page speed (separate from desktop)
– Content parity (is mobile missing content that desktop has?)
For each issue:
1. Show the specific HTML/CSS causing the problem
2. Explain how it affects mobile users
3. Provide exact code fix
4. Mention if it affects Google’s mobile-first indexing
Prioritize issues that actually break user experience, not minor design preferences.
When to use it:
After responsive design changes. If mobile traffic is high but conversions are low. When Mobile-Friendly Test shows warnings.
I used this for a client whose mobile traffic was 70% but conversions were 30% of desktop. ChatGPT identified tap targets too small (buttons were 32×32 pixels) and form fields requiring horizontal scrolling.
Fixed both issues. Mobile conversions increased 45% in two weeks.
Pro tip: Test actual mobile viewport, not just responsive desktop view. Mobile Chrome DevTools shows different results than actual device testing.
[VIDEO RECOMMENDATION: Screen recording showing how to paste Screaming Frog data into ChatGPT and interpret results]
How to Use These Prompts Effectively
Don’t just copy-paste and expect magic. Here’s what actually works:
Give ChatGPT context. The more data you provide, the better analysis you get. Paste in:
- Full HTML source code for markup checks
- Crawl data from Screaming Frog or Ahrefs
- PageSpeed Insights full report
- Search Console error reports
Be specific with your URL. Don’t analyze the homepage and assume it applies to all pages. Test different page types (product pages, blog posts, category pages).
Verify recommendations. ChatGPT sometimes suggests outdated fixes. Cross-check important recommendations with Google Search Central documentation.
Use follow-up prompts. If ChatGPT finds an issue, ask “How would this fix affect site speed?” or “Show me the exact robots.txt rule to implement this.”
Combine with real tools. These prompts work best alongside actual crawl tools. ChatGPT analyzes and prioritizes – it doesn’t replace Screaming Frog, Ahrefs, or GSC.
When These Prompts Don’t Work
Be realistic about limitations:
ChatGPT can’t crawl your site. You need to feed it data from actual crawl tools.
It sometimes suggests outdated practices. Always verify against current Google documentation.
It can’t see what Google sees. Use URL Inspection Tool in GSC to see rendered HTML.
It doesn’t have access to your analytics. Can’t tell you which technical issues actually hurt traffic.
Complex JavaScript rendering issues need real testing, not AI analysis.
Improving These Prompts
Make them better for your specific needs:
Add your tech stack. If you use WordPress, add “This is a WordPress site using [theme name]” to the prompt. ChatGPT will give WordPress-specific fixes.
Include your CMS limitations. If you can’t edit certain files, tell ChatGPT upfront so it suggests workarounds.
Specify your skill level. Add “Explain fixes in simple terms” or “Give me advanced technical implementation” based on your comfort with code.
Focus on business goals. Add “Prioritize fixes that improve conversion rates” or “Focus on issues affecting product pages” to get more relevant recommendations.
I’ve modified these prompts dozens of times based on client needs. The versions above are the ones that consistently find real issues.
They won’t replace proper technical SEO audits. But they’ll speed up initial diagnostics and help prioritize what to fix first.
