# ReAPI-com/mcp-openapi [Health: Active]

**Category:** 💻 Developer Tools  
**Repository:** https://github.com/ReAPI-com/mcp-openapi  
**GitHub Stars:** 90  
**npm Downloads (last month):** 156  
**Views:** 5  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/reapi-com-mcp-openapi

## Description
MCP server that lets LLMs know everything about your OpenAPI specifications to discover, explain and generate code/mock data

## Tools
Capabilities this server exposes over MCP:

- **refresh-api-catalog** — Refresh the API catalog
- **get-api-catalog** — Get the API catalog, the catalog contains metadata about all openapi specifications, their operations and schemas
- **search-api-operations** — Search for operations across specifications
- **search-api-schemas** — Search for schemas across specifications
- **load-api-operation-by-operationId** — Load an operation by operationId
- **load-api-operation-by-path-and-method** — Load an operation by path and method
- **load-api-schema-by-schemaName** — Load a schema by schemaName

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "mcp-openapi": {
    "command": "npx",
    "args": ["-y","@reapi/mcp-openapi@latest"]
  }
}
```

## Documentation

## What ReAPI-com/mcp-openapi MCP server does

ReAPI-com/mcp-openapi MCP server connects LLM-powered development tools to local OpenAPI 3.x specifications. It reads JSON, YAML, and YML files from a directory, processes their references into dereferenced schemas, and builds a catalog containing specification metadata, operations, and schemas.

The catalog gives an AI assistant a structured way to explore an API before writing integration code. It can identify available endpoints, explain operation parameters and responses, inspect data models, generate API-aware code, and produce mock data or request payloads based on the definitions in the specifications. These capabilities place the project in the 💻 Developer Tools category.

## How it works

At startup, the ReAPI-com/mcp-openapi MCP server scans the directory supplied through the command line. It processes each supported specification and creates an in-memory catalog. The `refresh-api-catalog` tool rescans the directory when files are added or changed.

MCP clients can retrieve the complete catalog or query it selectively. Searches can be limited to a specification ID, which is useful when several APIs contain similar paths, operation IDs, or schema names. A specification ID defaults to the source filename without its extension. An `x-spec-id` value in the OpenAPI `info` object can override that default and provide stable names for related specifications.

## Setup and configuration

Run the package through `npx` and pass the directory containing the specifications with `--dir`. Cursor can load the server from either a project-level `.cursor/mcp.json` file or a global `~/.cursor/mcp.json` file. A project-level configuration allows each project to expose a different group of specifications and helps keep catalog metadata within the model context limit.

The basic command uses `npx -y @reapi/mcp-openapi@latest --dir ./specs`. The target directory should contain OpenAPI 3.x files with `.json`, `.yaml`, or `.yml` extensions. After enabling the server in Cursor's MCP settings, use the refresh tool whenever the directory's specifications change.

## Tools and capabilities

- `refresh-api-catalog` rebuilds the catalog from the configured directory.
- `get-api-catalog` returns metadata for all loaded specifications, operations, and schemas.
- `search-api-operations` finds operations by query, optionally within one specification.
- `search-api-schemas` finds schemas by query, optionally within one specification.
- `load-api-operation-by-operationId` returns complete details for an operation ID and specification ID.
- `load-api-operation-by-path-and-method` loads an operation using its specification ID, endpoint path, and HTTP method.
- `load-api-schema-by-schemaName` returns a complete schema using its specification ID and schema name.

## Limitations and notes

The documented input is a local directory of OpenAPI 3.x files; remote specification synchronization is listed as future work. Semantic search and MCP-exposed code templates are also roadmap items, so the current search tools are catalog searches rather than the planned natural-language or template features. Multiple specifications with overlapping resource names should use distinct `x-spec-id` values to avoid ambiguity.

_Full upstream README: https://allmcps.com/mcp/reapi-com-mcp-openapi/readme_