openai/openai-mcpkit
Publicmirrored fromhttps://github.com/openai/openai-mcpkitAvailable
README.md
44lines · modecode
| 1 | # MCPKit: Secure MCP blueprints for enterprise data |
| 2 | |
| 3 | MCPKit is a blueprint for building authenticated Model Context Protocol (MCP) servers that let you bring proprietary data, content, and systems into ChatGPT, via [ChatGPT Dev Mode](https://platform.openai.com/docs/guides/developer-mode). |
| 4 | |
| 5 | Built by MCP engineers for enterprise builders, MCPKit accelerates the path from prototype to production when native connectors or existing MCP services do not cover your use case. |
| 6 | |
| 7 | ## Why build a MCP Server? |
| 8 | |
| 9 | - Serve high-value data directly inside ChatGPT through the connectors platform while respecting your existing authentication and entitlement rules. |
| 10 | - Keep sensitive content behind an authorization layer you control; MCPKit uses [Auth0](https://auth0.com/) as the example but supports any OIDC-compliant provider. |
| 11 | |
| 12 | ## Example use cases |
| 13 | |
| 14 | - **Financial services:** Expose internal research reports, expert-call transcripts, and other alternative data. The synthetic bundle in this repo mirrors alternative data feeds you can adapt to production. |
| 15 | - **Customer support & success:** Expose a gated knowledge base that blends CRM data, playbooks, and ticket summaries. |
| 16 | - **Healthcare & life sciences:** Expose documentation, SOPs, and clinical trial data. |
| 17 | - **E-commerce:** Expose item inventory, order status, fulfillment and delivery statuses. |
| 18 | |
| 19 | ## Sample data |
| 20 | |
| 21 | Use `synthetic_financial_data/` as a realistic sandbox for pipelines and demos. It contains alternative data artifacts such as analyst reports, expert-call summaries, and web-search trends with consistent tickers and timestamps. Swap in your own feeds once you are ready to plug into live systems. |
| 22 | |
| 23 | ## What you get in MCPKit |
| 24 | |
| 25 | - **Authenticated MCP server scaffolds:** [Python](python-authenticated-mcp-server-scaffold/README.md) and [TypeScript](typescript-authenticated-mcp-server-scaffold/README.md) servers that implement a number of different tools, include Deep Research-compatible `search` and `fetch` tools, apply entitlement checks, and follow the recommended resource/authorization separation model. |
| 26 | - **Authorization patterns:** The servers also implement the [MCP authorization specification](https://modelcontextprotocol.io/specification/2025-06-18/basic/authorization) pattern of separate resource and authorization servers. For the authorization server, we use an end-to-end Auth0 integration that you can replace with Okta, Azure AD, or an internal IdP by updating JWKS resolution, token validation, and tenant metadata. |
| 27 | - **Sample data:** A complete `synthetic_financial_data/` bundle with expert-call transcripts, pricing snapshots, and enrichment metadata so you can demo entitlement-aware tools before wiring up production feeds. |
| 28 | |
| 29 | ## Reference implementations |
| 30 | |
| 31 | - [TypeScript authenticated MCP server](typescript-authenticated-mcp-server-scaffold/README.md) |
| 32 | - [Python authenticated MCP server](python-authenticated-mcp-server-scaffold/README.md) |
| 33 | |
| 34 | Both implementations share a consistent API surface, emit structured logs, and lean on Auth0 for token exchange. Replace Auth0 with your preferred authorization provider by updating the server configuration documented in each README. |
| 35 | |
| 36 | ## Develop with ChatGPT Dev Mode |
| 37 | |
| 38 | - **Run locally:** Start either scaffold (`npm run dev` or `python -m server.app`) with your Auth0 application or alternate authorization server settings. |
| 39 | - **Expose securely via ngrok:** Tunnel the MCP server (`ngrok http <port>`) so ChatGPT can reach it during development without production deployment. |
| 40 | - **Register in ChatGPT Dev Mode:** Provide the tunneled URL, login with OAuth, and start querying. |
| 41 | |
| 42 |  |
| 43 | |
| 44 | - **Harden for production:** When you are ready, deploy your MCP server on your hosting platform of choice. |
| 45 | |