# agenthouse

> Scan any site for agent capabilities and payment protocols.

agenthouse produces Lighthouse-like reports for agentic services. Point it at any URL and it checks for:

- x402 / MPP / L402 payment protocols (HTTP 402 headers)
- Well-known discovery endpoints (/.well-known/x402.json, mpp.json, etc.)
- Bazaar service catalogs (discoverable agentic endpoints)
- MCP (Model Context Protocol) endpoints
- Claude and OpenAI plugin manifests
- OpenAPI specs, llms.txt, agents.txt, skills manifests
- Identity protocols (WebFinger, DID, Nostr, AT Protocol)

## API

GET  /api/scan?url=https://example.com — full JSON scan results (synchronous)
POST /api/scan {url} — queue async scan, returns job ID (requires database)
GET  /api/scan/:id — poll async scan job status and result
GET  /api/v1/scan?url=https://example.com — compact M2M scan (flat JSON, score)
GET  /api/badge?url=https://example.com — SVG badge for embedding
GET  /api/report?url=https://example.com — premium report with delta (x402 or MPP payment)
GET  /api/history?url=https://example.com — past scan summaries (requires database)
GET  /mcp — MCP tool manifest
POST /mcp — JSON-RPC 2.0 tool calls (scan_url, scan_compact, get_history)
GET  /health — health check
GET  /openapi.json — OpenAPI 3.0 spec

## Payment

agenthouse accepts x402 (Base/USDC) and MPP (Tempo/pathUSD) for premium endpoints.
Set X-Idempotency-Key header to prevent double-charges on retries.

## Example

curl "https://agenthouse.com/api/scan?url=https://agentnews.xyz"