Open source. Built for AI readiness.
Find what stands between AI and your website.
Agent Lighthouse is an open-source tool that checks how easily AI systems can find, understand, trust and use your site. Run your own scan. Get clear findings and practical next steps.
Free to use. Apache-2.0 licensed. Run it from your terminal or add it to your workflow.
For teams that depend on their website
Website owners · Marketing teams · SaaS products · Online shops · Documentation teams
From questions to next steps
See where your website needs work.
One scan brings the key signals together. Review the findings that fit your site and your goals.
AI discovery
Help AI find your pages
Find gaps in sitemaps, discovery files and links that help machines locate your content.
Explore the checksAnswers & citations
Make your answers easier to use
Check page structure, clear answers and supporting facts. See what could make your content harder to read or cite.
Explore the checksForms & actions
Make useful actions clear
Check forms and published action interfaces. Find missing labels or unclear instructions for AI tools.
Explore the checksCommerce readiness
Make your products easier to understand
Check product data, prices, availability and checkout signals. See where shop information needs attention.
Explore the checksCrawl access & safety
Set clear rules for access
Review crawler permissions and access barriers. Check secure connections and safety signals in the report.
Explore the checksResults & next steps
Know what to fix next
Get findings, priorities and guidance in a scan report. Start with relevant failures, then scan again after changes.
Explore the checks
Proof behind every check
Know what the result means.
Each check explains its rule, sources and limits. Only eligible checks with documented support affect the score. Advisory and experimental findings do not.
A scan measures signals on your site. It does not test every AI system or promise better search rankings.
Open source under Apache-2.0. Anyone can review the code.
Get started with Agent Lighthouse
Your next step starts with a scan.
Open-source project
Run Agent Lighthouse yourself.
Use your terminal to scan a site. Get an HTML report and a JSON file for your own workflow.
npx @forkpoint/agent-lighthouse https://yourstore.comFree and open source under Apache-2.0.
Another way to scan
Prefer to scan online?
Check your website with Agentic Storefront. Run an AI readiness scan in your browser without terminal setup.
Try Agentic StorefrontAn optional online alternative to running Agent Lighthouse yourself.
Explore what the scan checks
Go deeper into a topic. Each check explains what it measures and when it applies.
- Access & Crawl Control
- Content Extraction
- Machine Discovery
- Structured Data
- Answer Readiness
- Agent Interfaces
- Agentic Commerce
- Agent Operability & Safety
View a scan report
Already ran a command-line scan? Open its JSON file to see a score summary. Your file stays in your browser. Use the HTML report for full findings and fix guidance.
Share your result
Create a score badge for your README. Enter your scan result, then copy the snippet. This badge shows the score you enter; it does not verify a scan. Read the sharing guide.
[](https://github.com/ForkPoint/agent-lighthouse) <!-- Scanned: https://example.com -->
| Score | Colour | Meaning |
|---|---|---|
| 90–100 | 22c55e | Agent-ready |
| 70–89 | 4f46e5 | Good |
| 50–69 | f59e0b | Needs work |
| 0–49 | ef4444 | Blocked |
Have a public result? Share your score and what you changed.
Make your next website fix count.
Run Agent Lighthouse. See what needs attention.
Run your first scanFree and open source. Ready for your workflow.