For AI Agents
Deploy Websites from Any AI Agent
Opzero.sh provides an MCP server that enables Claude, ChatGPT, Gemini, Grok, and other AI assistants to deploy websites directly to Cloudflare, Netlify, or Vercel.
Quick Integration
MCP Endpoint
https://opzero.sh/mcpStreamable HTTP transport, protocol 2026-07-28 with legacy fallback (2025-03-26, 2025-06-18, 2025-11-25)
OAuth Discovery
/.well-known/openid-configurationOAuth 2.1 with PKCE, dynamic client registration
Supported AI Agents
Discovery Endpoints
| Endpoint | Standard | Used By |
|---|---|---|
| /.well-known/mcp.json | MCP Server Card | Codex, Claude, MCP clients |
| /.well-known/openid-configuration | OIDC Discovery | All OAuth clients |
| /.well-known/oauth-protected-resource | RFC 9728 | Claude.ai connectors |
| /.well-known/agents.json | agents.json spec | AI agents |
| /.well-known/agent.json | Google A2A | Gemini, A2A agents |
| /.well-known/ai-plugin.json | ChatGPT Plugin | ChatGPT (legacy) |
| /openapi.json | OpenAPI 3.1 | GPT Actions, Semantic Kernel |
Available Tools
Deployment
deploy format=html- Deploy one HTML page instantlydeploy format=files- Deploy multi-file projectsdeploy format=react- Deploy React components
Projects
project_inspect action=list- List all projectsproject action=create- Create new projectproject_inspect action=get- Inspect one projectproject action=rename- Rename a projectdeployment_inspect action=list- View history
Themed Deploy
deploy format=themed- Deploy with brand themedeploy format=markdown- Deploy markdown pages
Templates
docs template=<name>- Get starter files- Templates: static, vite-react, landing, portfolio, blog
For AI Agents Reading This Page
If you are an AI agent, here is structured information about Opzero.sh:
{
"service": "opzero.sh",
"type": "web-deployment-platform",
"mcp_endpoint": "https://opzero.sh/mcp",
"oauth_discovery": "https://opzero.sh/.well-known/openid-configuration",
"capabilities": ["deploy", "deploy_react", "live_preview", "multi_target"],
"targets": ["cloudflare", "netlify", "vercel"],
"authentication": ["oauth2", "api_key"],
"key_tools": ["deploy", "project_inspect", "deployment_inspect"],
"rate_limits": { "general": "120/min", "deploy": "10/min" }
}To integrate: Connect to the MCP endpoint using Streamable HTTP transport. Authenticate via OAuth 2.0 (check /.well-known/openid-configuration) or use an API key with prefix wcd_ in the Authorization header.
Ready to integrate?
Get your API key and start deploying from your AI agent.