anythingmcp

AnythingMCP

Turn any API, database or MCP server into custom connectors for Claude, ChatGPT and more — no code.

The self-hosted, AI-empowered MCP gateway that converts REST, SOAP/WSDL, GraphQL, SQL/NoSQL databases and other MCP servers into AI tools — then learns how their data connects and teaches your agents to use them, with auth and full audit.

 · 

 · 

AnythingMCP is a self-hosted, open-source smart, AI-empowered MCP gateway and MCP server that turns the systems you already run into Model Context Protocol tools — REST and SOAP APIs, GraphQL, SQL & NoSQL databases, and even other MCP servers. Import a spec or point it at a database, and expose it as a custom connector to Claude, ChatGPT, Gemini, Copilot, Cursor and any MCP-compatible client. No SDK, no code changes — point, configure, connect.

It ships with 188 ready-to-use adapters — including Deutsche Bahn, weclapp ERP, Etsy, Shopware, DHL and Sendcloud — so the most common integrations work in one click, while the visual editor and import tools (OpenAPI/Swagger, Postman, cURL, WSDL, GraphQL) let you wrap any other API or database in minutes.

What makes it smart, not just a pipe: AnythingMCP builds a per-workspace Knowledge Graph of how your connectors' data relates, serves it back to the agent over MCP so it chains tools correctly across systems, and turns how your tools are actually used into reusable AI skills. A plain gateway forwards calls; AnythingMCP gives your agents the context to use them well. *(All AI features are optional and opt-in — the gateway works fully without them.)*

And because every call runs through your own infrastructure, you decide what leaves it: per-tool response mapping declares which fields ever reach the model, so PII and secrets can be dropped before the answer goes out, while the audit log keeps the full upstream response on your side.

https://github.com/user-attachments/assets/2ae92f90-7012-4c00-8836-bae5a6422ca6

90-second demo — if the player doesn't load.

📖 Table of contents

• Get started in 60 seconds
• Key features
• Knowledge Graph & AI skills
• Control what the model sees
• Build custom Claude connectors — no code
• Turn your API into a ChatGPT app
• Why AnythingMCP
• Pre-configured MCP connectors
• Guides, client setup & FAQ
• Community & support
• Contributing
• License

Get started in 60 seconds

> 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 to `127.0.0.1` until you've created the admin account.

Or one-click deploy:

 

 

> Prefer manual setup? Copy `.env.example` to `.env` and run `docker compose up -d` — see the Deployment Guide.

Key features

• 5 connector types — REST, SOAP, GraphQL, Database (PostgreSQL, MySQL, MariaDB, MSSQL, Oracle, MongoDB, SQLite), MCP-to-MCP bridge
• 6 import formats + live discovery — OpenAPI/Swagger, Postman, cURL, WSDL, GraphQL introspection, custom JSON, plus tool discovery straight from a running MCP server
• 188 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
• **Knowledge Graph & AI skills** — a per-workspace, PII-safe map of how your connectors' data relates, served to the agent via an MCP tool, plus reusable AI skills composed into the server's instructions (optional, opt-in)
• **Response shaping & data governance** — declare per tool exactly which fields reach the model; drop PII, secrets and noise before they leave your network, with a live before/after preview
• Full auth — OAuth2 (PKCE + Client Credentials), Bearer, API Key, Basic, WS-Security, client certificates, LOGIN_TOKEN and OAuth 1.0a 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 up` and you're running

Knowledge Graph & AI skills

A gateway that only forwards calls leaves the hard part to the agent: knowing

which tool to call next, and what your business actually means by "open order"

or "active customer". AnythingMCP learns both — **how the data in your

connectors relates, and how your team really uses the tools** — then feeds

that back to the AI client as context rather than as extra tool calls.

• Knowledge Graph — a per-workspace map of *entities* (customers, orders,

products…) and their *relationships*. It builds itself from tool names,

parameters and the input/output of real calls; an optional AI pass infers the

cross-connector links heuristics miss. It stays PII-safe: it stores

entity/field *names* and relationship metadata, never the values.

• Build it visually — a graph editor lets you **create, edit and delete

entities and connections** by hand, add descriptions, and curate what the AI

proposed. Zoom/fit controls, connectivity-based layout and hover focus make a

large graph navigable.

• Served over MCP — each server exposes a `kg_how_to_obtain` tool so the

*customer's* agent can ask "how do I get this?" and receive chaining hints

across connectors.

• AI skills, written from real usage — with intent capture on, each tool

call can record *why* it was made. An AI pass turns recurring patterns into

small reusable rules (e.g. *"today's revenue includes order statuses 2, 3 and

4"*), scoped to a connector or a whole server. You Apply / Edit / Dismiss

each one, or let auto-apply take the high-confidence ones (≥ 0.90)

unattended. Applied skills are composed into the MCP server's instructions

at serve time, so they guide the agent without adding a single tool call,

and editing one takes effect on the next request. Consolidate with AI

merges overlapping rules back into a tight set as they accumulate.

The knowledge your team builds up by using the system stops living in

someone's head.

The AI passes (graph enrichment, skill generation, scheduled extension) work

with OpenAI, OpenRouter or Anthropic and are off by default — opt-in with a

global env flag *and* a per-workspace switch. The graph, manual editing and the

MCP tool work with no LLM key at all.

➡️ **Knowledge Graph & AI skills guide →**

Control what the model sees

Every tool can declare exactly which fields leave your infrastructure. The

mapping is attached per tool and applied on the way out, so the AI client — and

the third-party model behind it — only ever receives the shape you approved.

• Drop what should never travel. List the paths to remove and they are

stripped before the response reaches the agent: a customer's IBAN, an

employee's salary, an access token an API hands back alongside the data.

• Or declare the whole output. A `select` template names the fields to keep

and what to call them; a JMESPath expression covers the reshaping a template

can't express. Where an agent is better served by a stable shape, swap the

value for a placeholder (`"iban": "= [redacted]"`) instead of removing the

field.

• See it before you save it. The editor runs the mapping against a real

response and shows the before/after side by side, with the size difference. A

shipped adapter measures 12,172 B → 1,072 B (−91%) on a four-train result.

• Fails safe. A broken mapping returns the raw response and logs a warning

rather than breaking a working tool — unless you explicitly opt out.

Two payoffs at once: sensitive fields never reach the model, and every field you

drop is a field you don't pay for in the context window.

{
  "transform": {
    "mode": "select",
    "exclude": ["customer.iban", "customer.taxId"],
    "select": { "order": "$.id", "total": "$.amounts.gross", "status": "$.state" }
  }
}

> The audit log still records the full upstream response inside your own

> database. Shaping what the agent sees never costs you the evidence of what the

> API actually returned.

➡️ **Response mapping reference →**

Build custom Claude connectors — no code

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:

1. Import your API spec (OpenAPI/Swagger, Postman, cURL, WSDL, GraphQL introspection) or pick a pre-built adapter

2. Adjust tool names, descriptions and parameters in the visual editor — what the AI sees is up to you

3. 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 →

Turn your API into a ChatGPT app

Apps in ChatGPT — what OpenAI renamed connectors to in December 2025 — are built on MCP, and AnythingMCP gives you that 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 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 →

Why AnythingMCP

AI clients speak MCP, but your systems speak REST, SOAP, GraphQL and SQL. Writing and maintaining a bespoke MCP server per system — with auth, audit and access control — takes weeks each. AnythingMCP is the no-code layer in between:

Typical use cases — search train schedules and live delays with Deutsche Bahn · talk to your ERP from Claude (weclapp, 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.

Pre-configured MCP connectors

AnythingMCP ships with 188 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, in seven languages.

Guides, client setup & FAQ

Connecting an AI client, the connector types you can build, full documentation and the FAQ now live in one place:

➡️ **docs/guides.md** — Claude / ChatGPT / Gemini / Copilot / Cursor setup · REST / SOAP / GraphQL / Database / MCP-bridge connector guides · API reference & deployment docs · FAQ.

Looking for a specific service? Every adapter has a step-by-step guide at **anythingmcp.com/guides**.

Community & support

• 💬 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.

Contributing

We welcome contributions! Please read our Contributing guide before submitting a PR. For security issues, see SECURITY.md.

License

AnythingMCP is open source, licensed under the GNU Affero General Public License v3 (AGPL-3.0-only). Cloud-operator code under `ee/` directories is separately licensed and is not required for self-hosting — see the License FAQ.