liblab MCP Generator – API to MCP Server in 30 Seconds

TL;DR

TL;DR: liblab’s MCP Generator takes any OpenAPI or Postman spec, generates a hosted MCP server URL in 30 seconds, and lets you connect any REST API directly to Claude, Cursor, OpenAI, or Gemini as a native tool.

Source and Accuracy Notes

⚠️ This section is MANDATORY. All links must be verified from actual source, not guessed.

• Project page: mcp.liblab.com
• Source: No public source repository (commercial SaaS). License not applicable.
• HN launch thread: news.ycombinator.com/item?id=44379227 (17 points)
• Source last checked: 2026-07-07

What Is liblab MCP Generator?

Connecting a custom REST API to an AI assistant usually means writing a custom function-calling wrapper, handling auth, and maintaining it yourself. liblab MCP Generator automates that entire layer.

You drag-and-drop your API spec file, and liblab parses it, generates an MCP server, and gives you a hosted endpoint you can connect directly to any major AI tool. No code, no server to maintain.

Core features (verified from source):

• Accepts OpenAPI (Swagger) and Postman collection files as input
• Generates a hosted MCP server endpoint deployed via CDN worldwide
• Also offers a free downloadable local MCP server with unlimited calls
• Supports Claude, Cursor, OpenAI, and Gemini as target AI tools
• Privacy-first: no information about MCP requests is stored
• Pay-as-you-go pricing: $5 per 100 remote MCP calls; first 100 calls free per month

How It Works

Step 1: Bring Your API

Drag and drop your OpenAPI or Postman collection file into the liblab interface.

Step 2: Get Your MCP Server

Copy the remote MCP URL to share with your team or customers, or download a local MCP server binary for internal use. The local server is free with unlimited calls.

Step 3: Connect to Your AI Tool

Point your AI assistant (Claude, Cursor, OpenAI, or Gemini) at the MCP endpoint. Your API is now available as a native tool — the LLM can call your endpoints directly using natural language.

Pricing

| Plan | Details | |---|---| | Free tier | 100 remote MCP calls per month; free local server with unlimited calls | | Paid | $5 per 100 remote MCP calls; no setup fees, cancel any time |

Practical Evaluation Checklist

• Accepts OpenAPI and Postman specs — verified on product page
• Generates MCP server from spec in ~30 seconds — product claim
• Local MCP server option available — verified
• Privacy statement confirmed — “We don’t store any information about your MCP requests”
• Works with Claude, Cursor, OpenAI, and Gemini — verified from product page
• CDN-deployed remote endpoint — verified from product page

FAQ

Q: Does it support Swagger / OpenAPI 3.1? A: Yes. The product page explicitly lists OpenAPI and Postman as supported input formats.

Q: Is there a self-hosted option? A: Yes. A free local MCP server download is available with unlimited calls. No data leaves your infrastructure with the local option.

Q: What AI tools can connect to the generated MCP server? A: Claude, Cursor, OpenAI, and Gemini — verified from the product page.

Q: Does liblab store my API requests? A: No. The product states: “We don’t store any information about your MCP requests.”

Conclusion

liblab MCP Generator is a practical bridge between existing REST APIs and the growing MCP ecosystem. If you have an internal API you want to expose to AI assistants without building a custom function-calling layer, this fills that gap in minutes.

HN discussion was positive: users specifically called out how it removes friction compared to building your own function-calling integration. The local server option for privacy-sensitive use cases is a meaningful differentiator.

Source: mcp.liblab.com