Evidence-backed SEO fixes, delivered as pull requests
Find the fix.Ship the pull request.
Crawlmend is an AI SEO agent that crawls your site, reads your Search Console data, and opens a GitHub pull request for each fix — the file, the line, the before and after — for your team to review like any other change. It never injects a script into your site.
01 · Find
Evidence, not guesses
Every opportunity is grounded in a crawl finding or real Search Console performance — impressions, clicks, position — shown alongside the recommendation, not hidden behind it.
- Impressions
- 12,430
- Clicks
- 212
- CTR
- 1.7%
- Position
- 7.4
02 · Fix
A diff, not a suggestion
Proposed fixes arrive as an actual code diff against your repository — the file, the line, the before and after — so you're reviewing a change, not interpreting a paragraph.
03 · Ship
A pull request, not a merge
Approving a fix opens a branch and a pull request on GitHub. Nothing merges automatically — your team reviews it like any other change.
main from crawlmend/fix-142- Compile-checked before it opened
- Build passed
- Waiting for review from your team
Built for review, not autopilot
Fits the review process you already run
Crawlmend works alongside your existing GitHub workflow. It opens a dedicated branch and PR for each approved fix — your team reviews, comments, and merges exactly as they would any other contribution.
What changes in search
Same page. A result worth clicking.
Before
No meta description, so the search engine picks text from the page.
After the pull request
A title and description you wrote, reviewed and merged.
Illustrative example.
How it differs
Not another script on your site
Script-injection SEO tools
- Lives in a script that runs on your pages
- No review step: it changes the page after you ship
- The fix may only exist after JavaScript runs, and not every crawler runs it
- Stops working the day you remove the script
- Rolling back means finding what it changed
A pull request from Crawlmend
- Lives in your repo as an ordinary code change
- Reviewed by your team like any other change
- Already in the HTML you ship
- Stays after you stop using Crawlmend
- Rolling back is a git revert
Why that matters: JavaScript rendering is deferred and not guaranteed across search crawlers and AI bots, while a fix in your HTML is there for every one of them. Read the full comparison.
AI search
How Crawlmend helps you get cited by AI search
Answer engines can only cite pages they can fetch, read and understand. Crawlmend checks those mechanics on every crawl and opens a pull request for the fixes it can write. It can't guarantee a citation.
Crawlability for AI bots
Checks that answer bots such as PerplexityBot are allowed in robots.txt and not blocked at the CDN or firewall, and flags pages whose real content needs JavaScript to appear.
Structured data
Finds invalid schema and missing Organization, Breadcrumb and FAQPage markup, so systems can tell what a page is.
Answer-ready structure
Flags missing H2 sections, thin content and missing freshness dates on article pages, the things that make a passage quotable on its own.
llms.txt and sitemap
Reports a missing llms.txt or sitemap.xml. No major AI provider has confirmed it reads llms.txt, so it is a low-priority, low-cost fix.
A live visibility check
Runs a live web search for your queries and brand name and shows whether your domain appears among the sources. A sample, not a guarantee.
Read the developer checklist for getting cited by AI search.
What it works with
Sites whose code is on GitHub
Crawlmend needs a GitHub repository to open a pull request against. Here is what that covers today, stated plainly so you can tell quickly whether it fits.
Framework-aware fixes
Next.js (App Router), Nuxt, SvelteKit, Remix and React Router v7, Gatsby, Astro, Angular.
Detected, generic fixes
WordPress, Shopify, Hugo, Jekyll, Eleventy, Docusaurus, Django, Flask, Laravel, Symfony, Rails and more. Expect more review effort.
Compile-checked where possible
Repos with a tsconfig get a before-and-after type check; PHP, Python and Ruby edits get a syntax check; other repos get a lenient pass, not a guarantee.
Not supported yet
Sites with no code repository (hosted builders), and repositories on GitLab or Bitbucket.
Full detail in supported setups.
FAQ
Questions we get before someone connects a repo
Does Crawlmend ever push directly to my site?
No. Crawlmend never touches your production site or codebase directly. Every fix is opened as a pull request on GitHub for your team to review and merge, exactly like any other change.
Where does the evidence for a fix come from?
Every opportunity is grounded in a crawl finding plus real Google Search Console and PageSpeed data — impressions, clicks, position, and performance — shown alongside the recommendation.
What happens if I don't like a suggested fix?
Close the pull request. Nothing merges automatically and nothing changes on your site until your team approves and merges it.
How is Crawlmend different from SEO tools that inject JavaScript?
Script-injection SEO tools fix a site by adding a JavaScript snippet that edits pages in the browser, so the fix only exists once that script runs and is removed when you stop using the tool. Crawlmend instead opens a real, reviewable GitHub pull request — the file, the line, the before and after — so the fix lives in your code and in the HTML you ship.
Which frameworks does Crawlmend support?
Crawlmend has framework-aware fixes for Next.js (App Router), Nuxt, SvelteKit, Remix and React Router v7, Gatsby, Astro and Angular. It also recognises WordPress, Shopify, Hugo, Jekyll, Eleventy, Docusaurus, Django, Flask, Laravel, Symfony, Rails and other stacks, but fixes for those are generic and need more review. It requires the site's source code to be in a GitHub repository.
Does Crawlmend work with sites that aren't built with Next.js?
Yes, for sites whose code is in a GitHub repository. Next.js is one of seven frameworks with framework-aware fixes; other stacks are detected and get generic fixes. Sites with no code repository, such as hosted builders where you can't edit source files, and repositories on GitLab or Bitbucket are not supported yet: Crawlmend can't open a pull request without a GitHub repository to open it against.
What data does Crawlmend read?
Crawlmend reads your public pages, robots.txt, sitemap.xml and llms.txt; read-only Google Search Console performance data (queries, pages, impressions, clicks, click-through rate and position); public PageSpeed results; and the repository files it needs to detect your framework and write a fix. Page content is sent to OpenAI only when you trigger an AI-assisted action.
What permissions does Crawlmend request on GitHub?
Crawlmend is a GitHub App, so you choose which repositories to install it on and it can reach only those. It can read and write repository contents and pull requests, to create a new branch, commit and open a draft pull request, and read check results. It does not merge pull requests and does not push to your default branch. Access to Search Console is read-only.
Can Crawlmend help my site get cited by AI search?
Crawlmend checks the technical things that decide whether AI answer engines can read your pages: whether answer bots are allowed in robots.txt and not blocked at the CDN, whether content depends on JavaScript, structured data, answer-ready page structure and llms.txt. It also runs a live web-search check on whether your domain appears. It cannot guarantee that any AI assistant will cite you.