MCP Hub
Back to servers

anythingmcp

Convert any API into an MCP server in minutes. REST, SOAP, GraphQL, Database, MCP Gateway. Self-hosted, source-available.

GitHub
Stars
9
Forks
2
Updated
Mar 30, 2026
Validated
Apr 1, 2026

AnythingMCP

Convert any API into an MCP server in minutes.
REST API to MCP • SOAP to MCP • GraphQL to MCP • Database to MCP • MCP Gateway & Middleware

GitHub Stars  License  Release  Docker Ready  TypeScript  NestJS  Next.js  PRs Welcome  Contributors Welcome  Last Commit  Release Date

⭐ Star this repo if you find it useful — it helps others discover AnythingMCP!

AnythingMCP Banner


Cloud & Deploy

Try on Cloud    Deploy on Railway    Install on DigitalOcean

Self-hosting instead? Run ./setup.sh for the interactive Docker setup. See Quick Start below.


AnythingMCP is a self-hosted, source-available MCP middleware that turns your existing APIs into MCP (Model Context Protocol) servers. Connect any API — REST, SOAP, GraphQL, databases, or other MCP servers — and expose them as tools to AI clients like Claude, ChatGPT, Gemini, Copilot, Cursor, and more.

No SDK. No code changes. Just point, configure, and connect.

Looking for an MCP gateway? AnythingMCP acts as a universal MCP proxy and API-to-MCP bridge — the missing middleware between your APIs and AI agents.

AnythingMCP Demo
Watch the demo video


🏭 Built for production — AnythingMCP was born from real-world needs at a German industrial group connecting 15+ legacy systems (ERP, CRM, IoT) to AI agents.


Get Started in 60 Seconds

git clone https://github.com/HelpCode-ai/anythingmcp.git
cd anythingmcp && ./setup.sh
# Open http://localhost:3000 — done!

See the full Quick Start below for detailed configuration options.


Use Cases

  • Talk to your ERP from Claude Desktop — connect SAP, Oracle, or any REST/SOAP ERP and query it conversationally
  • Let AI agents query your production database safely — read-only database connectors with audit logging
  • Bridge legacy SOAP services to modern AI workflows — automatic WSDL parsing, no code changes
  • Aggregate multiple MCP servers behind one gateway — MCP-to-MCP bridge for unified tool access
  • Import your Postman collection and get MCP tools instantly — zero-config API onboarding

Why AnythingMCP?

ProblemSolution
You have REST APIs but AI clients speak MCPREST API to MCP conversion with OpenAPI/Swagger import
You have legacy SOAP/WSDL servicesSOAP to MCP bridge with automatic WSDL parsing
You need to query databases from AI agentsDatabase to MCP with auto-generated query tools
You want one MCP gateway for all your APIsMCP middleware that aggregates multiple connectors
You need auth, audit logs, and role-based accessBuilt-in enterprise governance layer

How AnythingMCP Compares

FeatureAnythingMCPCustom MCP ServerOther Gateways
No-code setup✅ Visual editor❌ Write code⚠️ Config files
SOAP / WSDL support✅ Built-in❌ Manual❌ Rare
Database connectors✅ 7 engines❌ Build yourself⚠️ Limited
Visual tool editor
Auth & audit trail✅ OAuth2, RBAC, logs❌ DIY⚠️ Partial
Self-hosted or Cloud✅ Docker / Railway / DigitalOcean / Cloud⚠️ Often SaaS-only
Multi-client support✅ Claude, ChatGPT, Gemini, Copilot, Cursor⚠️ Varies

Key Features

  • 5 Connector TypesREST, SOAP, GraphQL, Database (PostgreSQL, MySQL, MariaDB, MSSQL, Oracle, MongoDB, SQLite), MCP-to-MCP Bridge
  • 6 Import Formats — OpenAPI/Swagger, Postman Collections, cURL commands, WSDL, GraphQL introspection, custom JSON
  • Dynamic MCP Server — Tools registered at runtime, no restart needed
  • Visual Tool Editor — Map parameters to path, query, body, headers visually
  • Database Auto-Tools — Schema introspection + dynamic query execution out of the box
  • Environment Variables — Per-connector {{VAR}} interpolation, hidden from AI
  • Full Auth — OAuth2 (PKCE + Client Credentials), Bearer Token, API Key, Basic Auth, WS-Security, Certificates
  • Audit Logging — Every tool invocation logged with input, output, duration, status
  • Roles & Access Control — Tool-level whitelisting per custom role
  • Per-User MCP API Keys — Individual keys with usage tracking
  • Docker Readydocker compose up and you're running

Pre-configured Connectors

AnythingMCP ships with a catalog of ready-to-use connectors that work out of the box — just provide your API credentials and activate them. Currently available:

ConnectorCategory
DHL TrackingLogistics
MFR FieldserviceField Service Management
DATEVAccounting / Tax
BillomatInvoicing
FastbillInvoicing
ScopevisioERP / Accounting
weclappERP / CRM
KenjoHR
PayonePayments
N26 Open BankingBanking
ImmobilienScout24Real Estate
TeamViewerRemote Access
BundesbankFinancial Data
DESTATIS GenesisStatistics
NINA WarnungEmergency Alerts

Want to add your own? Create a JSON adapter file in packages/backend/src/adapters/ (organized by region, e.g. de/), register it in catalog.ts, and it becomes available to all users. See the existing adapters for the expected format.


Quick Start

git clone https://github.com/HelpCode-ai/anythingmcp.git
cd anythingmcp
./setup.sh                 # Interactive setup — generates .env, starts Docker

The setup script configures everything interactively: deployment mode, domain/SSL, auth, email, Redis, and more. All secrets are auto-generated. First user to register becomes Admin.

What setup.sh handles:

  • Domain and HTTPS — for production domains, enables Caddy reverse proxy with automatic Let's Encrypt SSL
  • Secrets — generates JWT, encryption keys, and database passwords
  • MCP authentication mode — OAuth 2.0, API Key, or both
  • Optional SMTP and Redis configuration

Prefer manual setup? Copy .env.example to .env, edit the values, and run docker compose up -d. See the Deployment Guide for details.

ServiceDefault URL
Web UIhttp://localhost:3000 (or https://yourdomain.com with Caddy)
Backend APIhttp://localhost:4000
MCP Endpointhttp://localhost:4000/mcp
Swagger Docshttp://localhost:4000/api/docs

Next step: Create a connector, import your API spec, and connect your AI client. See the Connector Guides below.


Connect Your AI Client

AnythingMCP works with any MCP-compatible client. Follow the guide for your AI tool:

ClientGuideTransport
Claude DesktopSetup GuideStreamable HTTP
Claude CodeSetup GuideStreamable HTTP
ChatGPTSetup GuideStreamable HTTP
Google GeminiSetup GuideHTTP / SSE
GitHub CopilotSetup GuideStreamable HTTP
CursorSetup GuideStreamable HTTP
Any MCP ClientSetup GuideStreamable HTTP

Connector Guides

Each connector type has dedicated documentation with setup instructions, examples, and best practices:

ConnectorUse CaseDocs
RESTHTTP APIs, OpenAPI/Swagger, PostmanREST Connector Guide
SOAPWSDL web services, WCF, legacy enterprise APIsSOAP Connector Guide
GraphQLGraphQL endpoints with introspectionGraphQL Connector Guide
DatabasePostgreSQL, MySQL, MariaDB, MSSQL, Oracle, MongoDB, SQLiteDatabase Connector Guide
MCP BridgeAggregate multiple MCP servers into oneMCP Bridge Guide

Architecture

                        ┌─────────────────────────────────┐
  Claude Desktop ──────►│                                 │
  ChatGPT ─────────────►│         AnythingMCP             │──── REST APIs
  Gemini CLI ──────────►│      (MCP Middleware)            │──── SOAP Services
  GitHub Copilot ──────►│                                 │──── GraphQL Endpoints
  Cursor ──────────────►│   MCP Protocol (HTTP)           │──── PostgreSQL / MySQL / MSSQL / MongoDB / ...
  Any MCP Client ──────►│                                 │──── Other MCP Servers
                        └─────────────────────────────────┘
                          Caddy (optional) │ automatic HTTPS
                          Next.js UI + NestJS Backend
                          PostgreSQL  │  Redis (optional)

How it works:

  1. Create a Connector — Point to your API (REST base URL, WSDL endpoint, GraphQL URL, database connection string)
  2. Import or Define Tools — Auto-import from OpenAPI/Postman/WSDL/GraphQL or define manually
  3. Connect AI Clients — Point your MCP client to http://your-server:4000/mcp
  4. AI calls tools — AnythingMCP translates MCP tool calls into actual API requests and returns results

Documentation

TopicDescription
API ReferenceFull REST API for connectors, tools, auth, audit
Tool Definition FormatParameters, endpoint mapping, response mapping
Deployment GuideDocker, production setup, reverse proxy, env vars
AuthenticationOAuth2, JWT, API keys, MCP auth modes

Tech Stack

LayerTechnology
FrontendNext.js 16, React 19, Tailwind CSS v4
BackendNestJS 11, TypeScript
MCP@modelcontextprotocol/sdk, Streamable HTTP
DatabasePostgreSQL 17, Prisma 7
CacheRedis 7 (optional)
Reverse ProxyCaddy 2 (optional — automatic HTTPS via Let's Encrypt)
AuthJWT, OAuth2, AES-256-GCM
DeployDocker (single container for app) + Docker Compose

Development

The easiest way to set up local development:

./setup.sh    # Choose "Local development" mode
npm run dev

Or see the Deployment Guide for manual setup.


Community & Support


Contributing

We welcome contributions! Please read our Contributing Guide before submitting a PR.

For security issues, see SECURITY.md.


License

AnythingMCP is source-available under the Business Source License 1.1 (BSL-1.1). This is not an OSI-approved open-source license — see the License FAQ for a plain-language explanation.

  • Free for: internal use, personal use, development, testing, evaluation, academic use
  • Not permitted: offering as a commercial hosted service (SaaS) without a separate license
  • Change Date: 2030-03-04 — on this date the license automatically converts to Apache 2.0

For commercial licensing: info@helpcode.ai

Transparency note: AnythingMCP makes optional network calls to anythingmcp.com for license verification and email delivery when SMTP is not configured. No API credentials or tool invocation data is ever sent. See External Services for full details.

Copyright (c) 2026 helpcode.ai GmbH

Reviews

No reviews yet

Sign in to write a review