From SEO scanner to your SEO agent.
Scans weekly. Grades every fix against Search Console. Gets smarter across every site. Here's how it gets built.
Three pieces that compound.
Every scan, every ranking change, every graded fix sharpens the next recommendation.
Scans without being asked
Weekly auto-scans per site. Every run captures what changed while you shipped.
Search Console grades every fix
Real clicks, impressions, and position per URL, daily. Every fix labeled by what actually moved.
Patterns across every site
Fingerprints, frameworks, niches. Day one you get “sites like yours typically…”
What you get today
Live now.
Every piece below is shipped and running. Free to try.
Scan any URL in under a minute.
40 checks across SEO, performance, accessibility, and AI search. Every issue returns with a fix prompt you paste into Claude or Cursor.
acmeshop.com · 8 issues found
Missing primary keyword in title
Title tag says "Home" with no keyword or ranking signal.
Found on page
<title>Home</title>
Fix prompt
Open app/layout.tsx and update the title to include your primary keyword in the first 30 characters.
Memory that builds across every scan.
Each scan compares against history. Regressions are flagged, not buried. You see what's improving and what keeps coming back.
acmeshop.com
↑ Improving7 scans · 42 days tracked · 84 URLs in sitemap
Missing og:image on /pricing
2dH1 missing on homepage
4dCanonical mismatch on /blog
1wInsight: Canonical issue on /blog has resurfaced 3×. Likely a deploy regression.
An SEO expert inside Claude Code.
Connect SEOLint as an MCP server in two minutes. Claude gets your scan history, open issues, and site profile. Ask what to fix, why a page isn't ranking, what to ship next.
claude code
❯ Where should I focus my SEO effort this month?
● get_site_intelligence (MCP)
⎿ acmeshop.com · 7 scans · 42d tracked
● get_open_issues (MCP)
⎿ 3 open · 1 recurring · 1 regressed
Fix the canonical on /blog first.
It's regressed 3 times, likely your build process. Then the og:image on /pricing, which is new and easy.
Your biggest gap is content: zero comparison pages despite selling in a competitive category.
Content direction, not just audits.
From your sitemap, SEOLint spots pages that should exist but don't. Each comes with a target keyword, intent, and a brief for Claude.
Content gaps · 3 page ideas
Best hiking boots under $100
/blog/hiking-boots-budget
Merrell vs Salomon trail shoes
/vs/merrell-salomon
Lightweight hiking gear guide
/blog/lightweight-gear
Each suggestion includes a copy-paste brief for Claude or Cursor.
The SEO Agent takes shape
Shipping now.
Turning the scanner into an ongoing partner.
Ongoing scheduled scans.
Flip it on for any site and SEOLint registers weekly re-scans. Silent when nothing changed, email when something moves.
acmeshop.com
WeeklyEmails you only when something changes. No noise.
Google Search Console plugged in.
Connect right after your first scan. Daily sync pulls real clicks, impressions, and position per URL. Every issue becomes a prioritised call, backed by data.
Search Console
+47%Clicks (30d)
2,418
Impressions
84.2k
Avg position
8.3
Real ranking data per URL. No guessing.
Pull requests with the actual fix.
Connect a GitHub repo. For mechanical fixes like meta descriptions, alt text, schema, and sitemap, SEOLint opens a real PR. You review and merge. The agent just writes the diff. Install on GitHub →
acmeshop/site · #142
fix(seo): add meta descriptions to 3 product pages
Flagged in last 3 scans. Drafted descriptions based on page content and your existing voice. Each is 140–155 chars.
+ app/products/trail-shoes/page.tsx
+ app/products/waterproof/page.tsx
+ app/products/lightweight/page.tsx
The long vision
Your on-call SEO engineer.
Where this is heading. One product covering the full loop: audit, track, fix, verify.
The SEO Agent, the product you subscribe to.
Connect once. The agent runs on a schedule you set. Depending on access level it returns code suggestions you paste into Claude or full pull requests that land in your repo. Every week you get a brief: what moved, what shipped, what's next. Open the agent →
Weekly brief · April 15
Sent every Monday. Skip weeks with nothing to report.
/blog/hiking-boots moved from position 14 to 6
Opened #142. Meta descriptions on 3 product pages.
Impressions on /pricing dropping. Flagged for inspection.
This week: Ship the PR. Investigate /pricing CTR.
Grading fixes by what actually moved.
Each scan captures your Search Console numbers. The agent compares them over time and tells you which fixes moved impressions, CTR, and position, and which didn't. Grading is just the change, made visible. Open your memory →
/blog/hiking-boots
GradedTitle rewrite shipped Apr 10
Position
+6
Impressions
+183%
CTR
+2.4pp
One map. Every setup step, every thing the agent just did for you.
A single visual tab that doubles as startup guide and progress board. Connections (GSC, GitHub, MCP) show status. Agent activity (scans, issues, PRs) shows counts. Upcoming events (next scan, first grade, trial end) show dates. Smart nudges (no sitemap found, impressions dropping, PR waiting) keep things moving. Startup guide on day one, agent control panel every week after.
Here's what's happening, quietly.
What the agent is doing
Scanning /pricing LIVE
Rendering, checking LCP + CLS.
Wrote fix for missing alt text FIX
12 images on /blog/*.
CTR recovered on /docs/agent CLEARED
What you need to do
Gets smarter across every site it touches.
Every scan feeds the engine. Fix types, frameworks, niches, page roles, result deltas from GSC. Over time the recommendation engine compounds: fewer generic best-practice suggestions, more “this specific change moved pages like yours by 18% last month.”
Across 127 sites
Next.js SaaS sites gained +18% impressions after shortening H1s to under 55 chars.
Your site fits. Suggested: shorten H1 on /pricing and /features.
AI page creation, informed by what actually ranks.
Gap found in your sitemap. Brief drafted from the posts and patterns already ranking in your niche. Claude writes the page. Scanned on publish, tracked via Search Console. Every post teaches the engine what shape of content lands, so next month's brief starts smarter.
Brief informed by what ranks
/guides/core-web-vitals-for-next-js
Modeled on 8 top-ranking posts in your niche.
Competitor gap analysis.
See what your competitors rank for that you don't cover. Turn gaps into briefs. Decide what to build.
competitor.com
3 keywords they rank for. You don't cover.
next.js seo checklist
4.2k/mo
ahrefs alternative
2.1k/mo
mcp server tutorial
1.4k/mo
lighthouse vs psi
980/mo
3 briefs ready. Hand to Claude to draft.
FAQ
Common questions.
Answers to what people ask before signing up.
What is SEOLint?
SEOLint is an SEO agent that watches your site weekly, grades every fix against Google Search Console, and ships code fixes as pull requests. Think ESLint, but for SEO. It lives in the dashboard, as a CLI, and as an MCP server inside Claude Code or Cursor.
How is SEOLint different from Ahrefs, SEMrush, or Screaming Frog?
Traditional SEO tools surface problems and stop there. SEOLint closes the loop: it detects the issue, writes the fix, opens the pull request, and grades the result against your real Search Console data. Built for developers who ship code, not marketers who run audits.
Do I need Claude Code to use SEOLint?
No. The dashboard works on its own. The MCP integration with Claude Code or Cursor is optional, a faster workflow for developers already in the terminal. CLI, REST API, and a GitHub Action are all included.
How does the agent grade my SEO fixes?
Every scan captures clicks, impressions, position, and CTR per URL from Google Search Console. When a fix ships, the agent watches those numbers at 14, 28, and 90 days. You see which fixes moved the needle and which didn't. No vanity metrics.
Will the GitHub PR bot push to my main branch?
No. The bot opens pull requests against a dedicated bot branch. You review and merge at your own pace. The GitHub App requests pull-request write access only, never direct push to protected branches.
What frameworks does SEOLint support?
Any website works for scanning. Framework-specific PR fixes have first-class support for Next.js today, with Astro, SvelteKit, and Remix on the roadmap. JSON-LD, meta, canonical, sitemap, and robots.txt fixes are framework-agnostic.
What's included in the 7-day free trial?
Everything. One site under agent watch, weekly auto-scans, 1,000 on-demand scans per month, GSC grading, MCP, CLI, REST API, and the GitHub PR bot. $99 per month after day 7, cancel anytime. A credit card is required at signup to prevent abuse.
How often does the agent scan my site?
Once a week by default, plus any on-demand scans you trigger yourself. You only get an email when something changes: ranks move, a new issue appears, a fix ships, or impressions drop. Weeks with nothing to report stay quiet.
Does SEOLint work for non-technical founders?
The dashboard, audits, and fix prompts are readable for anyone. The automated PR shipping and MCP workflow are designed for developers. Non-technical users can still hand fix prompts to a developer or to Claude to apply.
The scanner works today. The agent is coming.
Start now. Everything above gets added to your account as it ships.