📌 Introduction
As a developer building and maintaining backend APIs, I’ve always found API testing to be one of the most repetitive and time-consuming tasks. Writing individual test cases, ensuring coverage, and keeping tests updated with every change can be exhausting.
But that changed when I discovered Keploy — an AI-powered testing tool that automatically generates test cases by observing real API traffic.
In this blog, I’ll share how I used Keploy in my Inventory Manager project, my journey from 0% to 100% API test coverage, and why I think AI will transform how we approach testing.
🛠️ The Problem with Manual API Testing
Before Keploy, testing looked like this:
Manually writing test cases using Jest or Supertest
Creating sample data, setting up mock servers
Writing assertions for every request and response
Updating tests with every API update
Even for a small CRUD app, this could take hours — and often left you with poor coverage or outdated tests.
🤖 Introducing Keploy
Keploy offers a different approach. Instead of writing test cases manually, it records your real API traffic and auto-generates tests from it — complete with assertions and mocks.
It acts as a proxy, intercepting API requests and responses to turn them into reusable tests that can be run anytime.
📦 Think of it as turning real user behavior into test cases. No code, no boilerplate.
✅ What I Built: Inventory Manager
For this, I used my own project – a Node.js + Express app called Inventory Manager, with basic CRUD functionality:
GET /api/items
POST /api/items
PUT /api/items/:id
DELETE /api/items/:id
This project also supports CSV/Excel import, light/dark mode, and a frontend built using vanilla HTML, CSS, and JS.
🧪 Testing with Keploy: Step-by-Step
1️⃣ Install Keploy
On Linux/macOS:
bash
Copy
Edit
curl -sL https://get.keploy.io | bash
On Windows, download the binary from Keploy Releases and add the path to environment variables.
2️⃣ Start Recording API Calls
Keploy needs to record your traffic. You can run:
bash
Copy
Edit
keploy record -c "npm start"
Then make API requests using curl, Thunder Client, or your frontend.
Example:
bash
Copy
Edit
curl -X POST http://localhost:5000/api/items \
-H "Content-Type: application/json" \
-d '{"name": "Chair", "quantity": 5}'
Keploy will generate YAML test files under the keploy/tests folder.
3️⃣ Run the Tests
Once the traffic is recorded, you can test your API:
bash
Copy
Edit
keploy test
Keploy will replay the captured requests and compare the actual response against the recorded one.
🔁 CI/CD Integration with GitHub Actions
I also added Keploy into my CI/CD pipeline using GitHub Actions. This means:
Every time I push code
Keploy tests run automatically
Any regression is caught immediately
Here's what I added in .github/workflows/api-tests.yml:
yaml
Copy
Edit
- name: Run Keploy Tests
run: |
keploy test --config keploy.yaml
✅ The pipeline now builds, runs tests, and fails on regressions.
📊 Results & Coverage
In just a few minutes of interacting with my own app:
All CRUD endpoints were covered
Mocks were auto-generated
Test cases were created without writing any code
API regressions (if any) were caught instantly
🌟 Why I Loved Using Keploy
🧠 No manual test writing
🚀 Instant test generation from OpenAPI or real traffic
📉 Detects breakage due to changes
⚙️ Seamless CI/CD integration
📊 Test coverage improved effortlessly
It’s like having a QA assistant running in the background.
✨ Final Thoughts
Keploy changed how I think about testing. Instead of spending hours writing tests manually, I now:
Record real traffic
Auto-generate tests
Integrate with CI/CD
Focus on building, not debugging
If you're building REST APIs and want reliable testing without the overhead, give Keploy a try. It feels like the future of API testing.
📸 Screenshots & Resources
✅ Keploy test report screenshot (Add this to your README)
📘 My GitHub Repo
🧪 Keploy Docs
💬 Your Turn
Have you tried AI-powered testing?
Want help setting up Keploy on your own project?
Drop a comment below — I'd love to connect!