Build and test an Intergalactic Banking API end to end. Master API testing, automation with Postman Agent Mode, and building MCP servers to track transactions between astronauts. 60 minutes. Hands-on. No LLM experience required.
What you'll accomplish in 60 minutes
Use Postman Agent Mode to accelerate API design, testing, and development with AI copilots always available.
Move confidently between code-first and collection-first workflows using AI assistance to bridge the gap.
Build and validate MCP-enabled services compatible with modern LLM tooling and autonomous AI agents.
Apply a Postman-centric workflow to modernize APIs for an AI-first future — techniques you can use immediately.
Everything you need — open these before starting
Complete this before the workshop begins
In Postman, create a Blank Workspace and name it using this format:
Fork [Do It Yourself] Intergalactic Bank API from the public workspace into your newly created workspace:
🔗 postman.com/devrel/ai-powered-api-mcp-bootcampCreate an Environment Variable file called Banking.local
Switch to your new environment using the Environment Selector dropdown in the top right corner.
Ensure the Collection is set as context, then send:
For all the requests in this collection, add https://template.postman-echo.com as an environment variable called baseUrl and update all the URLs in the collection to use {{baseUrl}}.
baseUrl variable has been automatically populated in the Variables tab.
apiKey value to sensitive.
Set the fromAccount request as context and send:
Programmatically add a post-response script that parses the response body, reads the accountId field, and saves its value as an environment variable called fromAccount in the Banking.local environment.
Set the toAccount request as context and send:
Programmatically add a post-response script that parses the response body, reads the accountId field, and saves its value as an environment variable called toAccount in the Banking.local environment.
Send both requests, then send GET List All Accounts to view the randomly generated accounts.
Set the Create new transaction request as context and send:
Programmatically add a post-response script that parses the response body, reads the transactionId field, and saves its value as an environment variable called transactionId in the Banking.local environment.
Send the request to see the ID populated. Now send GET Transaction by ID — you should see a 200 OK response. 🎉
Once your full API request workflow is producing expected happy-path responses, set the Collection as context and send:
Write tests for all the requests in this collection. Include only status code tests and response time tests.
Walk through the Collection Runner interface. Demonstrate how to execute all requests in sequence and review results in the Test Results panel. Then use Agent Mode:
Run all tests using the Collection Runner.
Set authentication type to Bearer Token and enter your API key.
Run this command in your terminal to register the Postman MCP Server with Claude Code:
claude mcp add --transport http postman \ https://mcp.postman.com/mcp \ --header "Authorization: Bearer <YOUR_API_KEY>"
Expected mcp.json configuration:
{
"mcpServers": {
"postman_mcp_server": {
"url": "https://mcp.postman.com/mcp",
"headers": {
"Authorization": "Bearer PMAK-YOUR-KEY-HERE"
}
}
}
}
Run claude mcp list or type /mcp in Claude to list registered MCP servers. Then verify with:
List all my Postman workspaces.
Get your workspace ID:
Get the workspace ID for "[WORKSHOP-WORKSPACE-NAME]".
Add tests to the collection:
Using the Postman MCP server, add tests to [workshop collection name]. For each request: 1. Add a status code check (200 OK) 2. Validate the response has a 'success' property set to true
Run the tests:
Run the "[workshop collection name]" collection and show me the test results.
You've built and tested an Intergalactic Banking API end-to-end with Postman Agent Mode and MCP. These techniques are ready for your real-world projects.