# Moltalyzer Agent Policy
# https://api.moltalyzer.xyz
# Contact: api@moltalyzer.xyz

## Identity
Name: Moltalyzer
Type: Intelligence API
Base-URL: https://api.moltalyzer.xyz
Website: https://moltalyzer.xyz

## Agent Discovery
Agent-Card: https://api.moltalyzer.xyz/.well-known/agent-card.json
Agent-Protocol: https://api.moltalyzer.xyz/.well-known/agent.json
Agent-Manifest: https://api.moltalyzer.xyz/.well-known/agent-manifest.json
OpenAPI: https://api.moltalyzer.xyz/openapi.json
MCP-Package: moltalyzer-mcp (npm)

## Rate Limits
Free-Tier: 1 request per 5 minutes per IP on /latest endpoints
Free-API-Key: 5 requests per day (x-api-key header)
Paid-Tier: No rate limit (x402 micropayments, USDC on Base)
Burst-Limit: 10 requests per second per IP (all endpoints)
Brief-Endpoints: 1 request per minute per IP

## Authentication Methods
Method-1: none (free /brief, /index, /latest endpoints)
Method-2: x402 (USDC on Base Mainnet, per-request, no account)
Method-3: api-key (x-api-key header, free tier, 5 calls/day)
x402-Facilitator: https://x402.dexter.cash
x402-Manifest: https://api.moltalyzer.xyz/.well-known/x402.json

## Allowed Behaviors
Allow: Automated polling of /index endpoints to detect new data
Allow: Caching responses locally (respect Cache-Control headers)
Allow: Integrating digest content into agent workflows
Allow: Redistributing digest summaries with attribution
Allow: Automated x402 payments for paid endpoints
Allow: MCP server integration via moltalyzer-mcp npm package

## Disallowed Behaviors
Disallow: Scraping at rates exceeding stated limits
Disallow: Circumventing x402 payment requirements
Disallow: Impersonating Moltalyzer content as your own without attribution
Disallow: Accessing /stats or internal analytics endpoints

## Data Freshness
Intelligence-Digest: every 4 hours
Moltbook-Digest: every 1 hour
GitHub-Digest: every 24 hours
Token-Signals: real-time
Polymarket-Signals: real-time
Pulse-Narrative: every 4 hours

## Preferred Integration
1. Poll /index endpoints to detect new data (lightweight, free)
2. Fetch /brief for summaries (free, structured JSON)
3. Fetch /latest for full content (free, rate-limited)
4. Use x402 for historical data and batch signals
5. Use MCP server for native Claude integration