Build custom connectors for Claude and ChatGPT from any API — no code.
The self-hosted MCP gateway that turns REST, SOAP/WSDL, GraphQL and SQL endpoints into AI tools, with auth and full audit.
Try on Cloud → · Watch 90-sec demo → · Setup guides →
AnythingMCP is a self-hosted, open-source MCP server and API gateway that turns your existing APIs into Model Context Protocol tools. Connect any API — REST, SOAP, GraphQL, databases, or other MCP servers — and expose it as a custom connector to Claude, ChatGPT, Gemini, Copilot, Cursor, and any other MCP-compatible client. No SDK. No code changes. Point, configure, connect.
demo.mp4
90-second demo — direct link if the player doesn't load.
📖 Table of contents
Requires Docker 24+,
bash,openssl. On macOS, start Docker Desktop first.
git clone https://github.com/HelpCode-ai/anythingmcp.git
cd anythingmcp && ./setup.sh
# When setup finishes, open http://localhost:3000 and register
# the first user — they automatically become the admin.The interactive setup handles everything: deployment mode, domain & HTTPS (automatic Let's Encrypt via Caddy), secrets, MCP auth mode, optional SMTP/Redis.
⚠️ Register immediately after setup. The first account to register becomes Admin. If your instance is reachable from the internet during setup, configure firewall rules or bind the UI to127.0.0.1until you've created the admin account.
| Service | Default URL |
|---|---|
| Web UI | http://localhost:3000 |
| MCP endpoint | http://localhost:4000/mcp |
| Swagger docs | http://localhost:4000/api/docs |
Or one-click deploy:
Prefer manual setup? Copy
.env.exampleto.envand rundocker compose up -d— see the Deployment Guide.
Claude supports custom connectors: remote MCP servers you add once in Settings → Connectors, and that work across Claude.ai, Claude Desktop and Claude Code. AnythingMCP creates that connector from any API you already have — without writing an MCP server:
- Import your API spec (OpenAPI/Swagger, Postman, cURL, WSDL, GraphQL introspection) or pick a pre-built adapter
- Adjust tool names, descriptions and parameters in the visual editor — what the AI sees is up to you
- Add the gateway URL to Claude as a custom connector (OAuth 2.0 supported out of the box)
Your credentials stay on your infrastructure (AES-256-GCM at rest), every tool call lands in the audit log, and role-based access controls which users see which tools. Step-by-step guide →
ChatGPT Apps and connectors are built on MCP — and AnythingMCP gives you the MCP backend without writing one. Point it at your REST, SOAP, GraphQL or database endpoint and you get a ChatGPT-ready connector: add it in ChatGPT's connector settings (or use it as the tool layer of an Apps SDK app) and ChatGPT can read and act on your business data.
The same connector works simultaneously in Claude, ChatGPT, Gemini, Copilot and Cursor — build once, connect everywhere. ChatGPT setup guide →
| Problem | Solution |
|---|---|
| You have REST APIs but AI clients speak MCP | REST → MCP conversion with OpenAPI / Swagger import |
| You have legacy SOAP/WSDL services | SOAP → MCP bridge with automatic WSDL parsing |
| You need to query databases from AI agents | DB → MCP with auto-generated query tools |
| You want one MCP gateway for all your APIs | MCP middleware that aggregates multiple connectors |
| You need an MCP server for DHL/DATEV/Weclapp/… | 175+ pre-built adapters — install in one click |
| You can't ship credentials to a SaaS gateway | Runs on your infrastructure — credentials AES-256-GCM at rest |
| You need auth, audit logs, and RBAC | Built-in OAuth2, audit log, and role-based access |
Typical use cases — talk to your ERP from Claude (Weclapp, DATEV, Xentral) · track parcels with AI (DHL, GLS) · validate invoices (VIES VAT, Handelsregister) · let agents query production databases safely · bridge legacy SOAP to modern AI · import a Postman collection and get MCP tools instantly.
| Feature | AnythingMCP | Custom MCP server | Hosted MCP gateways |
|---|---|---|---|
| No-code setup | ✅ Visual editor | ❌ Write code | |
| SOAP / WSDL support | ✅ Built-in | ❌ Manual | ❌ Rarely supported |
| Database connectors | ✅ 7 engines | ❌ Build yourself | |
| Auth & audit trail | ✅ OAuth2, RBAC, logs | ❌ DIY | |
| Where credentials live | ✅ Your infra (AES-256-GCM) | ✅ Your code | |
| Pre-built SaaS adapters | ✅ 175+ ready-to-use | ❌ Build each | |
| Multi-client support | ✅ Claude, ChatGPT, Gemini, Copilot, Cursor | ✅ |
- 5 connector types — REST, SOAP, GraphQL, Database (PostgreSQL, MySQL, MariaDB, MSSQL, Oracle, MongoDB, SQLite), MCP-to-MCP bridge
- 6 import formats — OpenAPI/Swagger, Postman, cURL, WSDL, GraphQL introspection, custom JSON
- 175+ pre-built adapters — logistics, ERP, HR, e-commerce, payments, public data — see catalog
- Visual tool editor — map parameters to path, query, body, headers; rename and describe tools for the AI
- Dynamic MCP server — tools registered at runtime, no restart
- Full auth — OAuth2 (PKCE + Client Credentials), Bearer, API Key, Basic, WS-Security, client certificates, LOGIN_TOKEN handshakes
- Audit logging — every tool call logged with input, output, duration, status
- Roles & access control — tool-level whitelisting per custom role, per-user MCP API keys
- Environment variables — per-connector
{{VAR}}interpolation, hidden from the AI - Docker ready —
docker compose upand you're running
AnythingMCP ships with 175+ ready-to-use adapters — provide your API credentials at import time and the tools become available immediately. Every adapter has a setup guide on anythingmcp.com/guides (English, German, Italian).
| Category | Examples |
|---|---|
| 📦 Logistics & shipping | DHL, DPD, GLS, Shipcloud, Sendcloud, Deutsche Bahn |
| 💼 ERP, accounting & invoicing | DATEV, Weclapp, Xentral, Scopevisio, Billomat, FastBill |
| 🛍️ E-commerce | Shopware 6, WooCommerce, Mercado Libre 🌎, ImmobilienScout24, Oxomi |
| 👥 HR & field service | Personio, HRWorks, Kenjo, MFR Mobile Field Report |
| 🏛️ Government & public data | VIES VAT, Handelsregister, UK Companies House 🇬🇧, DESTATIS, Bundesbank, OpenPLZ, NINA |
| 🏦 Banking & payments | N26, Wise 🇬🇧, PAYONE, Razorpay 🇮🇳, Paystack 🇳🇬 |
| 💬 Messaging & communication | WhatsApp, LINE 🇯🇵, TeamViewer |
| 🎮 Gaming & Web3 | Sorare |
| 🏗️ Construction & mapping | PlanRadar, HERE Geocoding |
| Client | Guide | Transport |
|---|---|---|
| Claude Desktop / Claude Code | Setup → | Streamable HTTP |
| ChatGPT | Setup → | Streamable HTTP |
| Google Gemini | Setup → | HTTP / SSE |
| GitHub Copilot | Setup → | Streamable HTTP |
| Cursor | Setup → | Streamable HTTP |
| Any MCP client | Setup → | Streamable HTTP |
| Connector | Use case | Docs |
|---|---|---|
| REST | HTTP APIs, OpenAPI/Swagger, Postman | Guide → |
| SOAP | WSDL web services, WCF, legacy enterprise APIs | Guide → |
| GraphQL | GraphQL endpoints with introspection | Guide → |
| Database | PostgreSQL, MySQL, MariaDB, MSSQL, Oracle, MongoDB, SQLite | Guide → |
| MCP Bridge | Aggregate multiple MCP servers into one | Guide → |
| LOGIN_TOKEN auth | APIs that POST credentials → return long-lived bearer | Guide → |
How do I create a custom connector for Claude?
Run AnythingMCP (self-hosted or Cloud), import your API spec or pick a pre-built adapter, then add the gateway URL in Claude under Settings → Connectors. No code required — the Claude guide walks through it in ~5 minutes.
Can I build a ChatGPT App from my existing API?
Yes — ChatGPT Apps and connectors are built on MCP, and AnythingMCP generates the MCP backend from your existing API. Add it as a connector in ChatGPT, or use it as the tool layer of an Apps SDK app. See Turn your API into a ChatGPT App.
Do I need to know how to code?
No. Importing an OpenAPI/Postman/WSDL spec, editing tools in the visual editor, and connecting Claude or ChatGPT are all point-and-click. Code only comes into play if you want to contribute a new adapter (a single JSON file) or self-host with custom infrastructure.
What is an MCP server?
An MCP server exposes tools to an AI agent over the Model Context Protocol — an open standard from Anthropic. Once connected, the AI can call those tools to read data, run queries, or perform actions on your behalf. AnythingMCP is a self-hosted MCP server that wraps your existing APIs so you don't have to write one from scratch.
How is AnythingMCP different from writing my own MCP server?
You don't write code. AnythingMCP imports your OpenAPI / Postman / WSDL spec (or you point it at a database) and generates the MCP tools automatically. You also get auth, audit logging, RBAC, and a visual editor on top — features that would take weeks to build per service.
Is AnythingMCP really open source?
Yes. AnythingMCP is licensed under the GNU AGPL v3, an OSI-approved open-source license — the same model as Twenty, Cal.com, Grafana and Plausible. You can read, fork, modify, self-host and even offer it as a service; the AGPL's network copyleft simply requires that modifications to the software stay open. The only exception is code under ee/ directories (cloud-operator features), which is commercially licensed and not needed for self-hosting. See the License FAQ.
Is AnythingMCP free?
Yes. AnythingMCP is free software under the AGPL v3 — free for internal company use, personal use, development, testing, evaluation and academic use. If you modify it and run it as a network service for others, the AGPL requires you to make your modified source available to those users. For commercial licensing without copyleft obligations: info@helpcode.ai.
Can I self-host?
Yes — it ships as a Docker image and runs on your own infrastructure. Run ./setup.sh or use the Railway and DigitalOcean one-click installs. There's also a managed Cloud version if you'd rather not run it yourself.
What about SOAP and WSDL?
Built-in. AnythingMCP automatically parses WSDL documents and generates one MCP tool per SOAP operation. Useful for legacy enterprise APIs (SAP, Oracle, .NET WCF, banking middleware) that no AI client speaks natively.
Is MCP dead now that agents use CLI tools?
No — but the question conflates two problems. CLI is the right call when the model already knows the tool from training (git, docker, kubectl), the agent is acting for the builder, and a CLI actually exists. MCP wins when you need per-user auth, scoped permissions, audit logs, multi-tenant isolation, or SaaS integrations without a CLI. The mature pattern in 2026 is hybrid: CLI for local/dev tools, MCP for SaaS / multi-tenant / compliance-bound integrations. Full decision matrix on anythingmcp.com/vs/cli.
Can the AI access my production database directly?
Yes, with safety. PostgreSQL, MySQL, MariaDB, MSSQL, Oracle, MongoDB and SQLite are supported. Each tool is whitelisted, every invocation is audit-logged, and you can scope a connector to read-only credentials. See the Database Connector Guide.
How is auth handled?
OAuth2 (PKCE + Client Credentials), Bearer Token, API Key, Basic Auth, query-parameter auth, WS-Security and TLS client certificates are all supported. Credentials are stored AES-256-GCM encrypted at rest. Per-user MCP API keys are issued on top so each AI client gets its own key with usage tracking.
| Topic | Description |
|---|---|
| API reference | Full REST API for connectors, tools, auth, audit |
| Tool definition format | Parameters, endpoint mapping, response mapping |
| Deployment guide | Docker, production setup, reverse proxy, env vars |
| Authentication | OAuth2, JWT, API keys, MCP auth modes |
Next.js 16 + React 19 (frontend) · NestJS 11 + TypeScript (backend) · @modelcontextprotocol/sdk with Streamable HTTP · PostgreSQL 17 + Prisma 7 · Redis 7 (optional) · Caddy 2 (optional, automatic HTTPS) · Docker Compose.
Local development: ./setup.sh (choose "Local development"), then npm run dev — or see the Deployment guide.
- 💬 Questions & discussions — GitHub Discussions — vote on the next adapter, share what you've built
- 🐛 Bugs / 💡 features — Issues · 🆘 SUPPORT.md
- 🏢 Built by helpcode.ai in Freiburg, Germany — AnythingMCP was extracted from a production system connecting AI agents to 15+ legacy systems (ERP, CRM, SOAP, on-prem databases) in a German industrial group, and open-sourced because the catalog grows faster as a community. AI-assisted development, human-reviewed: see AUTHORS.md.
⭐ Like what you see? Star this repo — every star helps another developer discover AnythingMCP.
We welcome contributions! Please read our Contributing guide before submitting a PR. For security issues, see SECURITY.md.
AnythingMCP is open source under the GNU Affero General Public License v3 (AGPL-3.0-only) — see the License FAQ for a plain-language explanation.
- ✅ Free for — internal use, personal use, development, testing, evaluation, academic use, self-hosting
- 🔄 Copyleft — if you modify AnythingMCP and offer it over a network, you must share your modified source with those users
- 🏢
ee/directories — cloud-operator code underee/is licensed under the AnythingMCP Commercial License and is not required for self-hosting
