Simple pricing
Start free. Scale when you're ready.
Everything runs on your Claude API key. No card required.
Dev tool
Free
$0 /mo
The full SEO linter on your own Claude API key. No card.
- Scan any site
- 40+ SEO, performance & AEO checks
- AI fix prompts for every issue
- Full memory & scan history
- CLI, Dashboard & MCP (full access)
- Runs on your own Claude API key
- Search Console + per-fix grading
No card. No commitment.
Early birdAutonomous agent
SEOLint Agent
$99$79/mo
Everything in Free, plus the agent that runs SEO for you.
- Weekly auto-scans on autopilot
- Claude AI included, BYOK to extend
- Watches 1 site end-to-end, paced by your AI credits
- AI next-action oracle: what to fix next
- GitHub PR bot ships mechanical fixes
- AI-drafted blog posts from Search Console winners
- Programmatic SEO pages: free tools, glossaries, moreSoon
- Interactive email agentSoon
No charge until day 8.
For teams & agencies
Custom
Let's talk
Multiple sites with volume pricing and custom support.
- Everything in SEOLint Agent
- Multiple sites
- Higher AI usage ceiling
- More tracked pages per site
- Custom integrations
We'll build a plan that fits.
Your data stays yours
Encrypted. Never used for training.
Bring your own Claude key
We never see your code or key.
Cancel anytime
No contracts. No surprises.
Prices in USD, excluding VAT · VAT added at checkout for EU customers · Pay in any currency · Cancel anytime
Why Agent is one flat price
You hired a coworker, not a meter.
Tiered pricing with scan credits makes sense for a tool you check. It is the wrong model for an agent that runs in the background. Your coworker does not charge you per email sent. SEOLint charges one flat fee per site watched, and the agent runs as much as it needs to catch what matters.
Common questions
Full FAQ →- What's the difference between Free and the Agent plan?
- Free is the dev tool: scanner, AI fix prompts, full memory, Google Search Console grading, the read-only MCP surface. You drive it, on your own Claude API key. Agent ($79/mo) adds the autonomous layer on top: weekly auto-scans, AI next-action oracle, MCP write tools, GitHub PR bot, email agent. AI calls on the Agent plan are covered for typical use. Rule of thumb: manual is Free, autonomous is Agent.
- Do I need an Anthropic API key?
- Free tier: yes, after the one-time welcome scan. Add a key in Settings (60 seconds at console.anthropic.com) to keep getting AI prompts on every scan. Agent plan: no, AI usage is covered out of the box for typical agent activity. If you have heavier or unpredictable usage and would rather pay Anthropic directly, connect your own key on the Agent plan too. Either path works. The scanner, memory, Search Console, and MCP read tools work without a key on every tier.
- How does the 7-day free trial work?
- On the Agent plan, enter your card, no charge. The agent starts watching your site immediately and runs its first auto-scan within 24 hours so you see what it finds. On day 8, you're charged $79 for the first month and every month after. Cancel before day 8 and you pay nothing. Cancel later and you keep access until the end of your billing period. The Free tier doesn't need a card. Sign up and scan.
- What happens to my scan data if I don't upgrade on day 8?
- Nothing is deleted. Your scan history, Search Console data, memory, and reports stay in your account. You lose the ongoing agent (weekly auto-scans, memory updates, new fix grading) and fall back to free-preview quota, but everything already captured stays accessible and exportable. Upgrade later and the site picks up where it left off. If you want everything wiped, Settings has a GDPR self-delete that removes the account entirely.
- Why do you require a card to start the Agent trial?
- Two reasons. First, it filters out bots and serial trial abusers so the infrastructure stays affordable for real users. Second, it means you do not have to remember to come back on day 8. If SEOLint is working for you, you keep it. If not, one click cancels. The Free tier has no card requirement since AI runs on your own Claude key.
- What counts as one site?
- One site = one domain or subdomain. So example.com counts as one, app.example.com counts as a separate site. You pick which pages on that site the agent watches — homepage plus the paths you care about (pricing, top blog posts, product pages, docs). Most marketing sites land at 10-30 watched pages, comfortably inside the AI usage covered on the Agent plan. Multi-site or very large sites go to the Custom plan.
- What do I get, exactly?
- The agent watches 1 site with weekly auto-scans, runs the next-action oracle to surface the highest-ROI fix, ships mechanical fixes as GitHub pull requests, and drafts blog posts grounded in your Search Console data and validated against live SERPs. You also get manual scans you can trigger from MCP, CLI, or the dashboard. AI usage is covered for typical agent activity (a typical user picks 10-15 pages and finishes the month with credits to spare). Power users who want to remove the ceiling entirely can connect their own Claude API key. Everything else (Search Console grading, memory, AI fix prompts, REST API, GitHub Actions) is included on every tier.
- How does the blog post drafting work?
- The agent reads your Search Console data to find content gaps you could realistically rank for, then probes live Google SERPs for each candidate keyword to check it's winnable: page one isn't dominated by high-authority incumbents like Ahrefs or HubSpot, the search intent matches a blog post, and it fits your ICP. Promising keywords show up as suggestions in your todo list. Approve one and the agent drafts a full post grounded in your site's voice and existing pages, ready for you to review before it ships.
- Is the PR agent live yet?
- Yes, live today. The PR bot opens pull requests for mechanical fixes (titles, meta descriptions, alt text, schema, sitemap) on connected GitHub repos. Per-task token budgets keep runaway tasks from spiraling.
- Are there fair-use limits?
- Yes, sensible ones so one user can't accidentally exhaust shared infrastructure. Per-task token budgets, per-day spend caps, and a monthly AI ceiling on the Agent plan that covers typical use. If you have heavier or unpredictable AI usage, connecting your own Claude key on the Agent plan removes the ceiling entirely. Heavy multi-site users go to Custom.
- What if my site has lots of pages?
- The agent only watches the pages you pick, not all pages automatically. Start with the homepage plus your highest-value pages (pricing, top blog posts, key landing pages, product pages, docs). Most marketing sites land at 10-30 pages, well inside what the included AI usage covers. If you need more, connect your own Claude key to remove the ceiling, or email us for the Custom plan with a higher allowance.
- Can I watch more than one site?
- Yes, through the Custom plan. Email us with the number of sites you want tracked and we set up volume pricing. Agencies and founders running multiple properties typically land here.
- Can I try it before paying?
- Yes. Sign up and run a manual scan on any URL, free. You see the same issues and AI fix prompts that paid users get. The paid plan adds the ongoing agent that watches your site, syncs with Search Console, and builds memory over time.
- How do I use it day to day?
- Three surfaces, pick whichever fits. (1) Dashboard at seolint.dev. (2) CLI in your terminal or CI pipeline. (3) MCP server inside Claude Code, Cursor, Windsurf, or Claude Desktop. All three hit the same data, the same agent, the same memory.
- Do I need to know SEO?
- No. Every issue comes with a plain-English explanation and an AI fix prompt you paste into Claude, Cursor, or any AI agent. When the PR agent ships, it opens the pull request for you to review.
- Can I cancel anytime?
- Yes. Cancel from your dashboard at any time. You keep access until the end of the billing period.
Let the agent take SEO off your plate.
Sign up, connect your site, and the agent starts watching.
Get started for freeQuestions about a plan? Email support@seolint.dev. We reply within 3 business days. Payments processed by Creem, our merchant of record.