internal-swagger-mcp
Enables AI agents to search and retrieve API documentation from an internal Swagger platform via MCP.
README
internal-swagger-mcp
Let AI agents query your internal Swagger platform's API docs via MCP.
This server talks to the internal Swagger management platform's private share endpoint (
/flow/swagger/share?uid=...), not a public OpenAPI URL.
Tools
| Tool | Purpose |
|---|---|
swagger_list_sources |
List all configured services and their cache status |
swagger_search_api |
Search APIs by keyword (filterable by method / service) |
swagger_get_api_detail |
View an API's full parameters and mock example |
swagger_refresh_cache |
Force-refresh the doc cache (default TTL is 30 minutes) |
Connecting MCP clients
Requires Node.js ≥ 18. Swagger sources are always supplied by the client — this server holds no configuration. Pass them in stdio mode via the SWAGGER_SOURCES env var or --sources-file, and in HTTP mode via the X-Swagger-Sources header per request. Use project scope for every client's MCP config so each repo pins its own sources and the config can be committed to git. In the snippets below, <SOURCE> looks like http://your-server/...#/swaggerManage?uid=xxx; if swagger_list_sources works inside the client, the integration is up.
Start in HTTP mode (for deploying on a shared internal host):
npx -y internal-swagger-mcp --http # defaults to port 3000; override with --port or PORT
Claude Code
Official docs — using --scope project writes to the project root's .mcp.json.
Local (stdio):
claude mcp add swagger --scope project --env SWAGGER_SOURCES='["<SOURCE>"]' -- npx -y internal-swagger-mcp
Remote (HTTP):
claude mcp add --transport http swagger --scope project http://<internal-IP>:3000/mcp --header 'X-Swagger-Sources: ["<SOURCE>"]'
opencode
Official docs — place this in opencode.json at the project root.
Local (stdio):
{
"mcp": {
"swagger": {
"type": "local",
"command": ["npx", "-y", "internal-swagger-mcp"],
"environment": {
"SWAGGER_SOURCES": "[\"<SOURCE>\"]"
}
}
}
}
Remote (HTTP):
{
"mcp": {
"swagger": {
"type": "remote",
"url": "http://<internal-IP>:3000/mcp",
"headers": {
"X-Swagger-Sources": "[\"<SOURCE>\"]"
}
}
}
}
Cursor
Official docs — place this in .cursor/mcp.json at the project root.
Local (stdio):
{
"mcpServers": {
"swagger": {
"command": "npx",
"args": ["-y", "internal-swagger-mcp"],
"env": {
"SWAGGER_SOURCES": "[\"<SOURCE>\"]"
}
}
}
}
Remote (HTTP):
{
"mcpServers": {
"swagger": {
"url": "http://<internal-IP>:3000/mcp",
"headers": {
"X-Swagger-Sources": "[\"<SOURCE>\"]"
}
}
}
}
Sources file
When the source list belongs to the project, pass --sources-file <path> instead of pasting the same JSON-as-string into every client's env. Use a path relative to the project root (e.g. ./swagger-sources.json) — it resolves from process.cwd(), which is the project root under project-scoped configs in Claude Code, Cursor, opencode, etc. — so the MCP config can be committed and shared as-is.
swagger-sources.json (each entry is a <SOURCE> URL as defined above):
[
"<SOURCE_1>",
"<SOURCE_2>"
]
Each client config then becomes a thin wrapper around the same command:
Claude Code:
claude mcp add swagger --scope project -- npx -y internal-swagger-mcp --sources-file ./swagger-sources.json
opencode (opencode.json):
{
"mcp": {
"swagger": {
"type": "local",
"command": ["npx", "-y", "internal-swagger-mcp", "--sources-file", "./swagger-sources.json"]
}
}
}
Cursor (.cursor/mcp.json) — and other clients using the mcpServers shape:
{
"mcpServers": {
"swagger": {
"command": "npx",
"args": ["-y", "internal-swagger-mcp", "--sources-file", "./swagger-sources.json"]
}
}
}
The file is read once at startup; the source list is fixed for the server's lifetime (clients relaunch on config change anyway). When both --sources-file and SWAGGER_SOURCES are provided, the file wins. The flag is rejected in --http mode because HTTP sources are inherently per-request.
HTTP deployment security
The server binds to 0.0.0.0 by default for easy intranet sharing, and prints a warning if started bare. In production, set at least one of the following:
| Environment variable | Effect |
|---|---|
MCP_BIND_HOST |
Bind address; set to 127.0.0.1 to restrict access to the local host (default 0.0.0.0) |
MCP_BEARER_TOKEN |
Require an Authorization: Bearer <token> header on every request |
MCP_ALLOWED_ORIGINS |
Comma-separated Origin allowlist (DNS-rebinding protection) |
When
MCP_ALLOWED_ORIGINSis set, requests without anOriginheader are rejected — except for requests carrying a validMCP_BEARER_TOKEN, so server-to-server calls still work.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.