Real Crawlmend pull requests, step by step
Two fixes from crawl finding to merged pull request, on a public demo repository.
Example 1: a title that is too short
The finding
| Value | |
|---|---|
| Issue | Title tag too short |
| Page | /services/web-design |
| Found by | Site crawl |
| Evidence | This title tag is only 8 characters: "Services". |
| Why it matters | A very short title leaves most of the space Google shows in search results unused, wasting a chance to include relevant keywords or a clear value proposition. |
With Search Console connected, Crawlmend would add the page's real impressions, clicks and position next to this finding, so you could see whether the page already earns the traffic that makes the fix worth doing.
What changes
| Title tag | |
|---|---|
| Before | Services |
| After | Fixed-Scope Next.js Web Design for Studios |
The new title was written by an AI model from the page's own content, because padding a short title mechanically would read as automated. It is a suggestion for your team to accept, edit or reject. Crawlmend does not promise that it will change rankings or clicks.
The diff
Crawlmend detected a Next.js project, mapped the URL to its source file, and edited the exported metadata object.
export const metadata = {
- title: "Services",
+ title: "Fixed-Scope Next.js Web Design for Studios",
};Verification
Before showing the diff, Crawlmend applied it to a throwaway clone of the repository and ran its TypeScript check. The page showed "Compile-checked: this diff was verified to build clean before you saw it". See what compile-checked means for what this does and does not cover.
The pull request
| Field | What Crawlmend created |
|---|---|
| Title | Crawlmend fixes — /services/web-design |
| State | Opened as a draft, then marked ready for review once CI passed |
| Branch | crawlmend/e0f6d2ae0ef2eb2f, a new branch from the default branch |
| Commit | One commit: "Title tag too short — /services/web-design" |
| Changes | 1 file, 1 line added, 1 line removed |
| Description | The issue, why it matters, the recommendation and what was applied |
| Author | The GitHub account that connected Crawlmend |
Open pull request #1 on GitHub to see the real diff, commit and description.
Example 2: a missing sitemap
The finding
| Value | |
|---|---|
| Issue | Missing sitemap.xml |
| Impact | High |
| Found by | Site crawl |
| Evidence | No sitemap found at https://crawlmend-demo-site.vercel.app/sitemap.xml. |
| Why it matters | Without a sitemap, search engines rely entirely on discovering pages through internal links, so new or deep pages can take much longer to get indexed, or be missed entirely. |
The diff
This one needed no AI call. Crawlmend built the sitemap from the four pages its crawl had found and proposed it as a new file in the framework's public/ folder, where Next.js serves it at /sitemap.xml.
+<?xml version="1.0" encoding="UTF-8"?>
+<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
+ <url><loc>https://crawlmend-demo-site.vercel.app/</loc></url>
+ <url><loc>https://crawlmend-demo-site.vercel.app/services/web-design</loc></url>
+ <url><loc>https://crawlmend-demo-site.vercel.app/contact</loc></url>
+ <url><loc>https://crawlmend-demo-site.vercel.app/about</loc></url>
+</urlset>Verification
The diff showed "Not yet compile-checked", and the approve button stayed disabled until the check was run. After the check passed ("Compile-checked: this diff was verified to build clean before you saw it"), the pull request could be opened.
The pull request
| Field | What Crawlmend created |
|---|---|
| Title | Crawlmend fixes — / |
| State | Opened as a draft, then marked ready for review once CI passed |
| Branch | crawlmend/42099b4af021e53f |
| Commit | One commit: "Missing sitemap.xml — /" |
| Changes | 1 new file, 7 lines |
Open pull request #2 on GitHub.
After the merge
- It was merged once the Vercel preview deploy had succeeded. Crawlmend itself never merges: a person does.
- The production deploy then served
/sitemap.xmlwith the four pages (HTTP 200,application/xml). - Crawlmend marked the pull request as Merged when its Pull Requests page next loaded: it has no webhooks, so it checks GitHub on page load. A new crawl no longer reported the missing sitemap.
Try it on your own site
The same flow works on your repository. Read how fixes are generated for the details, or connect a repository to see your own findings.