MCP Link logo

MCP Link

by automation-ai-labs

Converts OpenAPI V3 specifications into MCP interfaces for AI agents without changing the original API.

Install MCP Link

Pick your client, then add it in one click or paste the config.

No confirmed install command yet

We only publish a config when the install details come from the project itself — its README, its docs, or a verified owner. We haven’t found those for automation-ai-labs/mcp-link, and we’d rather show nothing than a guess you’d paste into your client.

Open setup steps on GitHub

Overview

The automation-ai-labs/mcp-link MCP server converts OpenAPI V3 specifications into MCP tools that AI agents can call. It runs as a Go service, forwarding tool requests to a target REST API using a configured base URL and authentication header. Endpoint filters can restrict which operations are exposed. Reach for it when an API already has an OpenAPI description and needs an MCP interface without manually writing a wrapper.

Use cases

  • Expose an OpenAPI-described REST API to MCP clients
  • Generate agent tools from an existing API schema
  • Filter generated tools by path or HTTP method
  • Test an API-to-MCP conversion with the hosted service

Key features

  • OpenAPI V3 to MCP conversion
  • REST API tool forwarding
  • Configurable authentication headers
  • Path and method endpoint filters
  • Go-based local service
  • Hosted conversion and testing endpoint

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked Today

Recent health check succeeded.

20/20 checks healthy over the last 84d

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

GitHub stars625Stargazers on the source repository.
Last commit1y agoMost recent push to the default branch.
Directory activity3 viewsConfig copies, upvotes, and views on AllMCPs.

How MCP Link works

What automation-ai-labs/mcp-link MCP server does

The automation-ai-labs/mcp-link MCP server turns an OpenAPI V3 document into an MCP-compatible interface. It maps the operations described by the specification to tools that AI agents can invoke, allowing an existing REST API to participate in MCP-based workflows without changes to the API implementation.

The repository lists examples including Brave Search, DuckDuckGo, Figma, GitHub, Home Assistant, Notion, Slack, Stripe, TMDB, and YouTube.

How it works

MCP Link runs as a Go application that serves an MCP endpoint. A request identifies the OpenAPI specification with s= and the target API base URL with u=. The service reads the API definition, constructs the corresponding MCP tool interface, and sends calls to the configured API.

Authentication can be supplied through h= using the format header-name:value-prefix. This supports configurations such as API keys, API tokens, and bearer tokens when they can be represented by an HTTP header.

Endpoint selection is controlled with f= path filters. A filter can include or exclude a path, select an HTTP method, or combine multiple rules with semicolons. Single-segment and multi-segment wildcards are supported, so a deployment can expose only the required portion of an API.

Setup and configuration

Clone the repository, enter the project directory, and download its Go dependencies with go mod download. Start the service with go run main.go serve, optionally providing --port and --host, for example port 8080 and host 0.0.0.0.

Configure an MCP client with a URL shaped like:

text
http://localhost:8080/sse?s=[OpenAPI-Spec-URL]&u=[API-Base-URL]&h=[Auth-Header]:[Value-Prefix]

Replace the bracketed values with URL-encoded configuration for the specification, API endpoint, and authentication header. Add f= when endpoint filtering is needed. A hosted version is also available at mcp-link.vercel.app for converting and testing APIs without a local installation.

Tools and capabilities

The automation-ai-labs/mcp-link MCP server provides capabilities derived from the selected OpenAPI schema rather than a fixed catalog of tools. Its supported functions include:

  • Converting OpenAPI V3 operations into MCP tools
  • Forwarding agent calls to a configured REST API
  • Passing authentication headers to the target API
  • Including or excluding endpoints with path and method filters
  • Serving the resulting interface over an MCP connection

Limitations and notes

The generated tool surface depends on the completeness and compatibility of the OpenAPI specification. The project describes OAuth support for the MCP protocol, resource handling, and broader MIME-type support as future development items, so these should not be assumed to be available.

The configuration examples use SSE and require the API specification URL, target base URL, and any needed authentication details. Review the selected endpoints and credentials before exposing the service to an agent. The project is listed in the Developer Tools category and is most relevant when an existing API already has an OpenAPI V3 schema.

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about MCP Link

automation-ai-labs/mcp-link is an MCP server that converts OpenAPI V3 specifications into MCP interfaces for REST APIs. It exposes API operations as agent-callable tools, using configurable base URLs, authentication headers, and endpoint filters. Developers can run it locally or use the hosted version to convert and test APIs.

Clone the repository, enter its project directory, and run `go mod download` to install Go dependencies. Start it with `go run main.go serve`, adding `--port` and `--host` options as needed.

Connect through the service's SSE URL, supplying the OpenAPI specification with `s=`, the target API base URL with `u=`, and authentication details with `h=`. The README provides a JSON MCP configuration using a URL such as `http://localhost:8080/sse?...`.

Yes, use the `f=` parameter to include or exclude paths and HTTP methods. Filters support `*` for one path segment, `**` for multiple segments, and semicolon-separated rules such as including GET operations while excluding an internal path.

Related MCP servers

Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewMCP Link AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/automation-ai-labs-mcp-link?style=directory)](https://allmcps.com/mcp/automation-ai-labs-mcp-link)
HTML Embed
<a href="https://allmcps.com/mcp/automation-ai-labs-mcp-link"><img src="https://allmcps.com/api/badge/automation-ai-labs-mcp-link?style=directory" alt="MCP Link on AllMCPs" /></a>