Obenan MCP Server — AI-Native Business Management via ChatGPT & Claude (47 Tools, Production)
Waseem Mansha
Published
What It Is
A production MCP (Model Context Protocol) server that connects AI apps like ChatGPT and Claude directly to the Obenan platform — letting business owners manage their locations, reviews, and entire online presence through natural language conversations.
Live: https://www.twine.net/signin
What It Does
Everything you do from the Obenan dashboard — now accessible directly from ChatGPT or Claude:
- Manage business locations and profiles
- Analyze and reply to customer reviews
- Track Google Search performance
- Update business hours, descriptions, keywords, contacts, and menus
- 47 tools, all through natural conversation
No dashboards. No clicking through menus. Business owners interact with their data the way they think — in plain language.
How It Works
- Connect ChatGPT or Claude to
mcp.obenan.aiand sign in with your Obenan account - The AI assistant auto-discovers all 47 available tools
- Ask anything — “How are my reviews this month?” or “Update my business hours”
- The server securely calls the Obenan API on your behalf and returns results
- Your AI assistant presents everything in a clear, conversational format
Built With
| Layer | Technology |
|---|---|
| Runtime | Python 3.13 — FastMCP (FastAPI), async HTTP, structlog |
| Auth | OAuth 2.0 + PKCE with JWT validation |
| Deployment | Docker + AWS ECS Fargate via Terraform (IaC) |
| CI/CD | GitHub Actions — automated deployments on every push |
| Testing | 1,730 unit + integration tests with real API validation |
Why It Matters
This project sits at the intersection of AI agent tooling and real-world SaaS integration — one of the most in-demand engineering skills in 2025. It demonstrates:
- Deep expertise in the Model Context Protocol (MCP) — the emerging standard for AI tool integration
- Production-grade OAuth 2.0 + PKCE security for multi-tenant AI access
- Infrastructure as Code on AWS with full CI/CD automation
- A complete test suite of 1,730 tests ensuring production reliability
If you need an engineer who can build AI-native integrations, agentic backends, or MCP servers for your product — this is what that looks like in production.