# Postman's MCP Server Now Works With Google Antigravity IDE

We are excited to announce that Postman's MCP server now integrates natively with Google's Antigravity IDE. If you are already using Postman to manage your APIs, you can now give Antigravity's AI agents direct access to your Collections, environments, and test workflows with a few lines of config.

## **What Is Postman's MCP Server?**

 The[ Postman MCP server](https://www.postman.com/product/mcp-server/) uses the Model Context Protocol (MCP) to give AI agents structured, executable access to your APIs. Think of it as a translation layer between your Postman workspace and any MCP-compatible AI tool. You define your APIs once in Postman, and the server handles the rest. Once connected, AI agents can browse your Collections, execute requests, run your test scripts, and reference your environment variables to move between dev, staging, and production. Your Collections stop being documentation. They become something an AI can actually use. ## **Setting Up the Integration**

 To get started, generate a Postman API key at[ go.postman.co/settings/me/api-keys](https://go.postman.co/settings/me/api-keys). ![](https://blog.postman.com/wp-content/uploads/2026/04/install.gif)1. Open the MCP store via the "Additional options” or “...”dropdown at the top of the editor's agent panel.
2. Click on "MCP Servers"
3. Search “Postman”
4. Click “Install”
5. Enter your Postman API Key
 
 That's it. Antigravity's AI agent will automatically discover the Collections and environments in your workspace. Your APIs show up as callable tools in the agent's context, and you're off and building. ![](https://blog.postman.com/wp-content/uploads/2026/04/2-configured-1024x690.png)## **How It Works Under the Hood**

 When the Postman MCP server starts up, it reads your connected Postman workspace and exposes your Collections as tools the AI agent can invoke directly. Each request in a Collection becomes a callable action. Your environments become switchable contexts. Your test scripts become validation steps the agent can run and report on. Antigravity maintains this context persistently across sessions. The agent doesn't start from scratch each time you open a new task. It already knows your API surface, and it can draw on that knowledge throughout a full feature build without you having to re-explain anything. ## **What You Can Do With It**

 Rather than share a long feature list, here are a few real usage scenarios. **"My team needs to validate API changes before merging"** This is one of the most common pain points for API teams. A backend change ships, something breaks downstream, and no one finds out until the PR is already merged. With the integration active, you can ask Antigravity's agent to run your Postman test suite against staging before any pull request is opened. It executes the Collection, parses the results, surfaces failures inline, and flags which requests broke. Regressions get caught before they reach code review. **"I'm writing code that calls an internal API and I don't want to guess at the contract"** This happens constantly. A developer writes an integration against a service they don't own, and they're either digging through static docs or copying examples from memory. With Postman's MCP server connected, Antigravity's agent pulls the actual request definition from your Collection. The correct method, headers, auth scheme, and expected response shape are all there. No hallucinated endpoints. No made-up field names. The agent is working from what the API actually does. **"We onboarded a new engineer and they need to understand our API surface"** Point them to Antigravity and have them ask the agent to walk through your APIs. The agent can execute live requests against a sandbox environment and show real responses in context. It is a faster, more concrete introduction than handing someone a static doc and hoping they read it. ![](https://blog.postman.com/wp-content/uploads/2026/04/3-newemployee-1024x689.png) **"I have an OpenAPI spec and I need a Postman Collection"** Drop the spec into Antigravity and ask the agent to scaffold a Collection from it. It uses Postman's MCP tools to generate requests, test assertions, and example responses, then publishes the Collection directly to your workspace. Your team can pick it up immediately. ## **Why This Is Different From Just Using AI With Your Docs**

 Most AI coding tools can read a markdown file or an OpenAPI spec. What they cannot do is run your API, check whether the endpoint is actually up, or tell you whether the response coming back from staging today matches what the spec says it should return. Postman's MCP server gives the agent grounded, executable context. It is the difference between an AI that talks about your APIs and one that can actually call them. ## **Get Started**

 The Postman MCP server is available now. The Antigravity IDE integration is live for all Postman users today. Check out our[ MCP server documentation](https://learning.postman.com/docs/postman-ai/mcp-server/) for a full walkthrough, or explore the[ open-source repo on GitHub](https://github.com/postmanlabs/postman-mcp-server). If you have questions or want to share what you're building, head over to the[ Postman Community](https://community.postman.com/). *Have questions about Postman's MCP server? Visit our*[ *product page*](https://www.postman.com/product/mcp-server/) *or join the discussion in the*[ *Postman Community Forums*](https://community.postman.com/)*.*