The 5-minute checklist
These four files make your site visible to AI agents. Each takes minutes to add and immediately improves your agentgrade score.
1. Add llms.txt
Create /llms.txt — a plain-text description of your service for LLMs:
# Your Service Name
> One-line description of what your service does.
## API
GET /api/search?q=term — Search items
POST /api/create — Create an item (requires auth)
## Example
curl "https://yoursite.com/api/search?q=test"
Why: LLMs read this to understand what your service does and how to use it. Start the file with an H1 heading for a verified check.
2. Add OpenAPI spec
Serve /openapi.json describing your API endpoints:
{
"openapi": "3.0.0",
"info": { "title": "Your API", "version": "1.0" },
"paths": {
"/api/search": {
"get": {
"summary": "Search items",
"parameters": [
{ "name": "q", "in": "query", "schema": { "type": "string" } }
]
}
}
}
}
Why: Agents use this to know exactly what endpoints exist and how to call them. Most frameworks can auto-generate this.
3. Add robots.txt with agent directives
Add agent-specific rules to /robots.txt:
User-agent: GPTBot
Allow: /
User-agent: Claude-Web
Allow: /
User-agent: Anthropic-AI
Allow: /
Why: Signals that your site welcomes AI agents. Many sites block AI crawlers by default.
4. Add a health endpoint
Serve /health returning JSON:
{ "status": "ok" }
Why: Agents and monitoring tools use this to check if your service is up before making requests.
Next steps
Once the basics are in place, consider:
- MCP endpoint — Let AI models call your tools directly via JSON-RPC 2.0 (learn more)
- Payment support — Accept micropayments from agents via x402 or MPP (learn more)
- Service catalog — Publish discoverable endpoints via Bazaar (learn more)
- Skills manifest — Curate the key actions agents should take (learn more)
Scan your site
Enter your domain on agentgrade.com to see your current score and what's missing.