# Cloudpylon auth.md

> Authentication & Access Policy for AI Agents and Automated Systems.

## 1. Overview & Authentication Model

Cloudpylon is a free, open-access multi-cloud compute pricing and specification database.

- **Authentication Requirement**: **None (Public & Open Access)**
- **API Keys / Tokens**: **Not Required**
- **User Accounts / Registration**: **Not Required**
- **Base Website**: https://cloudpylon.com
- **Public API Base**: https://api.cloudpylon.com/v1
- **Web App API Proxy**: https://cloudpylon.com/api

Autonomous AI agents, LLM tool runners, search crawlers, and automated pipelines can query all catalog and pricing data directly without registering, obtaining API keys, or passing bearer tokens.

---

## 2. Available Public Endpoints

All data endpoints are publicly queryable:

### All Compute Instances
- **Endpoint**: `GET https://api.cloudpylon.com/v1/instances` (or `GET https://cloudpylon.com/api/instances`)
- **Format**: JSON
- **Filters**: `provider`, `min_vcpu`, `max_vcpu`, `min_memory`, `max_memory`, `regions`, `sort`, `order`

### Cheapest & Best Value Rankings
- **Endpoint**: `GET https://cloudpylon.com/api/instances/cheapest`
- **Format**: JSON

### Provider Native Endpoints
- **Endpoint**: `GET https://cloudpylon.com/api/{provider}/instances`
- **Supported Providers**: `aws`, `azure`, `gcp`, `hetzner`, `oci`, `digitalocean`, `linode`, `alibaba`, `tencent`, `ovhcloud`, `netcup`, `scaleway`, `atlanticcloud`

---

## 3. Machine-Readable Discovery Files

Agents can discover complete API documentation and catalog summaries via:

1. **RFC 9727 API Catalog**: `https://cloudpylon.com/.well-known/api-catalog`
2. **OpenAPI Specification**: `https://api.cloudpylon.com/docs/openapi.json`
3. **Interactive API Documentation**: `https://api.cloudpylon.com/docs`
4. **LLMs Discovery Index**: `https://cloudpylon.com/llms.txt`
5. **Full Technical LLM Reference**: `https://cloudpylon.com/llms-full.txt`

---

## 4. Rate Limiting & Usage Guidelines

While authentication is not required, agents should follow standard good-citizen guidelines:
- Identify your agent using a descriptive `User-Agent` header.
- Respect HTTP caching headers (`Cache-Control`, `ETag`).
- For bulk queries, utilize pagination parameters (`limit`, `page`).