MCP Gateway
A gateway to translate MCP tool calls into HTTP API requests, configurable via YAML.
MCP Gateway
MCP (Model Context Protocol) Gateway can translate MCP tool callings to traditional HTTP API requests. It can provide a configurable way to get existing HTTP API to MCP territory.
Getting Started
Create config file from config.example.yaml:
$ cp config.example.yaml config.yaml
Edit config.yaml file, map all APIs to MCP tools.
Then start launch it with SSE transport:
$ uv run mcp-gateway
INFO: Started server process [15400]
INFO: Waiting for application startup.
INFO: Application startup complete.
INFO: Uvicorn running on http://0.0.0.0:3001 (Press CTRL+C to quit)
Default is 3001.
Server Control
Change Port
Provide parameter --port=<port_no> in command line will change the port to SSE transport.
Launch gateway with port 3002:
$ uv run mcp-gateway --port=3002
INFO: Started server process [15400]
INFO: Waiting for application startup.
INFO: Application startup complete.
INFO: Uvicorn running on http://0.0.0.0:3002 (Press CTRL+C to quit)
stdio Transport
Provide parameter --transport=stdio in command line will change the transport to stdio.
E.G.:
$ uv run mcp-gateway --transport=stdio
It's meaningless to manually launch gateway in stdio transport. You can configure in Cursor or Cline like:
{
"mcpServers": {
"mcp-gateway": {
"command": "uv",
"args": ["run", "mcp-gateway", "--transport=stdio"]
}
}
}
Or MCP Inspector with form values:
| Item | Value |
|---|---|
| Transport Type | STDIO |
| Command | uv |
| Arguments | run mcp-gateway --transport=stdio |
Configuration File
There is two parts in configuration YAML, server and tools. server defines the basic info for gateway server use. tools defines the mapping from single MCP tool to HTTP API request.
server:
name: rest-amap-server
config:
apiKey: foo
tools:
- name: maps-geo
description: "将详细的结构化地址转换为经纬度坐标。支持对地标性名胜景区、建筑物名称解析为经纬度坐标"
args:
- name: address
description: "待解析的结构化地址信息"
required: true
- name: city
description: "指定查询的城市"
required: false
requestTemplate:
url: "https://restapi.amap.com/v3/geocode/geo?key={{.config.apiKey}}&address={{.args.address}}&city={{.args.city}}&source=ts_mcp"
method: GET
headers:
- key: x-api-key
value: "{{.config.apiKey}}"
- key: Content-Type
value: application/json
responseTemplate:
body: |
# 地理编码信息
{{- range $index, $geo := .Geocodes }}
## 地点 {{add $index 1}}
- **国家**: {{ $geo.Country }}
- **省份**: {{ $geo.Province }}
- **城市**: {{ $geo.City }}
- **城市代码**: {{ $geo.Citycode }}
- **区/县**: {{ $geo.District }}
- **街道**: {{ $geo.Street }}
- **门牌号**: {{ $geo.Number }}
- **行政编码**: {{ $geo.Adcode }}
- **坐标**: {{ $geo.Location }}
- **级别**: {{ $geo.Level }}
{{- end }}
Server
| Item | Description |
|---|---|
| name | Server name |
| config | Key/Value pairs that can be referenced by var {{.config.xxx}} in templates |
Tools
tools is list of MCP tools mapping. Single tool props. are defined as follows:
| Item | Description |
|---|---|
| name | Tool name (Function name), which is provided to LLM. |
| description | Tool description through which LLM can understand what the tool could do. |
| args | Arguments of tool (Function arguments). |
| requestTemplate | Request mapping to target HTTP API. |
| responseTemplate | Response mapping for response of target HTTP API. |
Single argument props. are defined as follows:
| Item | Type | Description |
|---|---|---|
| name | Argument name, which is provided to LLM. | |
| description | Argument description through which LLM can understand and decide what value should be filled. | |
| required | Boolean | Required argument or not. |
Request template props. are defined as follows:
| Item | Description |
|---|---|
| method | HTTP method |
| url | Target HTTP API url template |
| headers | HTTP headers |
HTTP headers are defined as follows:
| Item | Description |
|---|---|
| key | Header key |
| value | Header value template |
Response template props are defined as follows:
| Item | Description |
|---|---|
| body | Response body template |
Contribution
All kinds of contribution are welcomed.
Related Servers
Scout Monitoring MCP
sponsorPut performance and error data directly in the hands of your AI assistant.
Alpha Vantage MCP Server
sponsorAccess financial market data: realtime & historical stock, ETF, options, forex, crypto, commodities, fundamentals, technical indicators, & more
MCP Time Server
Provides tools for getting the current time and date, and formatting timestamps in various formats and timezones.
Text Classification (Model2Vec)
A server for text classification using static embeddings from Model2Vec, supporting multiple transports like stdio and HTTP/SSE.
MCP-Haskell
A complete Model Context Protocol (MCP) implementation for Haskell, supporting both StdIO and HTTP transport.
@mcp-fe/react-tools
Don't let AI guess from screenshots. Give LLMs direct access to your React state, Context, and Data Grids. Features bidirectional communication via SharedWorkers & WebSockets. Docker gateway included.
SynapseForge
A server for systematic AI experimentation and prompt A/B testing.
Solana Metrics MCP Server
Analyze Solana metrics from InfluxDB and generate Grafana dashboards.
CIE - Code Intelligence Engine
Local code analysis MCP server with 25+ tools: semantic search, call graph tracing, dependency analysis, and symbol navigation. Built with Tree-sitter and CozoDB. Supports Go, Python, JS, TS.
Tuteliq
AI-powered safety tools for detecting grooming, bullying, threats, and harmful interactions in conversations. The server integrates Tuteliq’s behavioral risk detection API via the Model Context Protocol (MCP), enabling AI assistants to analyze interaction patterns rather than relying on keyword moderation. Use cases include platform safety, chat moderation, child protection, and compliance with regulations such as the EU Digital Services Act (DSA), COPPA, and KOSA.
Grafana
Search dashboards, investigate incidents and query datasources in your Grafana instance
nf-core MCP Server
Manage and navigate local nf-core pipeline repositories.