Qodo: an AI copilot for coding and testing
At Codana, we write a lot of code, so we're always looking for ways to make our processes run smoother and deliver even better digital experiences. We recently came across Qodo: a tool that uses Artificial Intelligence (AI) to help us code and test. In this blog post, you can read all about our initial experiences.
What is Qodo?
Simply put, Qodo is an AI tool that helps developers. It's a kind of testing agent that helps to automate and improve software tests, and also lends a hand in generating code and streamlining testing workflows.
With classic test automation, you have to manually write and maintain test cases yourself, but Qodo uses AI to devise, analyse, and improve those scenarios on its own.
So, what can it do?
- Qodo looks at your existing code and functions and suggests test cases based on them.
- The AI helps to spot errors and potential weak points faster, which makes the software better.
- Qodo fits within modern DevOps workflows and can be linked to CI/CD pipelines. Important for how we like to handle development.
- Last but not least: you get reports and suggestions to boost the software's quality.
Qodo in practice
If you talk the talk, you’ve got to walk the walk, so we decided to try out Qodo for ourselves. There are plug-ins for VS Code and JetBrains, so our developers could just stay in their familiar IDEs. We tried out the VS Code integration.
What immediately stood out was the AI chat feature. Especially when you're just starting on a project, it's handy. Just ask Qodo "How do I run this project?" or "What is this project about?", and it gives you a clear explanation of how the project is structured and what you need to do to get it running locally.
That saves you plenty of time digging through documentation, so an AI assistant makes it much easier for new team members (or colleagues jumping in) to get up to speed quickly.
How does Qodo generate tests?
Qodo makes generating tests a lot simpler by automatically suggesting test scenarios and the corresponding code. A quick way is the "Test this function" option right in your code. Qodo then creates tests specifically for that function, matching its logic.
You can also ask for tests for specific files via the general AI Gen chat. Qodo then asks some targeted questions (via a sort of wizard) to know exactly what you need. Based on that, you get custom test scenarios. Once you’re ready, you can save them to a test file with one click. It fits neatly into your project and helps further automate your testing workflow.
What's also good: Qodo thinks beyond the standard 'happy flows'. The AI also comes up with edge cases: precisely those things you sometimes miss as a developer, but which are important for truly stable, reliable software.
How do we use Qodo?
We primarily see Qodo fitting into the types of projects we often do at Codana: custom web applications, mobile apps, and more complex systems.
Because the tool isn't tied to one programming language (it's language-agnostic), we can use it for various technologies like PHP, Laravel, JavaScript, and React Native. That matches our approach, as we often work with both front-end and back-end code, mobile apps, and APIs.
For teams that, like us, work agile and value DevOps, AI testing like Qodo can help get feedback faster and deliver more stable software.
Qodo helps... but the developer decides!
Important to remember: Qodo is a good assistant, but not a replacement for a "real" developer. The tool automates a lot, but our developers always retain control. They are the ones who ultimately review, adjust, and decide on all of Qodo's suggestions.
It's a tool that supports, and that's exactly how we like to see technology work. It helps us not only to work faster but, more importantly, to build even better and more reliable digital experiences that truly make an impact and make users happy. We're definitely keeping an eye on it!
