CanAgentUse

Scanner for the agentic web

Is your website ready for AI agents?

Scan your site to see if AI agents can find, crawl, and use it.CanAgentUse runs a public, evidence-based scan of your homepage and discovery endpoints. It checks whether agents can find machine-readable files, crawl useful content, understand API and MCP contracts, respect bot policies, and return actionable remediation for GEO, AIO, AEO, and Lighthouse SEO readiness.

Scanner coverage

129 signals across 15 readiness areas for the AI agent age.

The scanner measures whether websites and web apps give AI systems the signals they need to discover public pages, trust crawler policy, understand content, call documented APIs, and complete approved workflows.

Agent reliability

66

Discovery, content, APIs, auth, MCP, commerce

Security & policy

9

Headers, crawler trust, training policy

GEO, AIO and AEO

22

AI citations, AI Overview, and answer-engine readiness

Performance & accessibility

32

Performance, accessibility, SEO, and best practices

Latest guides

Agent-readiness field notes

View all posts

Developer Toolkit

Integrate the MCP Server. Automate CLI in CI/CD.

The readiness scanner operates as a standalone CLI tool, an MCP server, and a CI/CD regression gate. Connect it to Claude or Cursor for local repair workflows, run public or localhost scans from the terminal, and compare saved reports to block regressions before a preview build ships.

Integration Options
Step-by-step Setup

Connect to Claude Desktop or Cursor

Configure your AI editor to run the scanner locally. This allows the model to analyze directories, fetch playbooks, and draft config files automatically.

1. Copy Server ConfigurationShared configuration format
{
  "mcpServers": {
    "can-agent-use": {
      "command": "npx",
      "args": ["-y", "can-agent-use@latest"]
    }
  }
}
2. Insert into config files:

Claude Desktop

Paste configuration into:

~/Library/Application Support/Claude/claude_desktop_config.json

Cursor IDE

Settings → Features → MCP → Add Server (or edit):

~/.cursor/mcp.json

💡 Try asking your assistant:

›"Scan https://localhost:3000 and fix failing readiness rules"
›"Compare our production site with local build and show regressions"
›"Create a validated robots.txt file that permits AI bots"
Stateless & Fast
Local-First Safe
Signals
Playbook-Linked
Fully open source on GitLab under MIT License
View source on GitLab·Install: npm i -g can-agent-use

Connect with us

Want help making your site agent-ready?

Send your site and goals. We can review readiness gaps, implementation options, and the right next step before or after you scan.

Meeting link unavailable

FAQs

Questions about AI-agent readiness

The scanner focuses on the signals agents need to discover your site, understand its content, and use available interfaces reliably.

It checks whether AI agents can discover, access, and use your site by reviewing signals like robots.txt, sitemap.xml, llms.txt, OpenAPI, API catalog, MCP metadata, structured data, bot access rules, and core page quality.

AI agents depend on clear machine-readable signals, crawl permissions, semantic content, and usable interfaces. Improving those signals helps agents understand your site and complete useful tasks without guessing.

No. Enter a public website URL and the scanner runs from the web. For the most accurate results, scan a production page that does not require a login.

You get a report with an overall readiness score, category scores, evidence for each check, and recommended fixes for making the site easier for AI agents and search systems to use.

Teams responsible for growth, product, engineering, ecommerce, and SEO can use it to find gaps in agent discoverability, API discovery, structured data, and web quality.