Testeranto
The AI-powered BDD test framework for polyglot projects.
Vibe Coding meets
the Real World
You've heard the hype - now it's time to get down to business. Harness the raw power of LLMs with the control of BDD, and you can vibe out like a pro.
Tests so smart
they fix themselves
When tests fail, an LLM can automatically resolve the problem with a single click. They handle type errors and static analysis too. Informed by your documentation, they can even implement new features. You might never write code by hand again.
Context is key
Testeranto maximizes the LLM's context by including only the most relevant source files, documentation, and test results. This means you can get better results from smaller models. The context is so well-structured that you can even run OpenRouter models locally and avoid paying those pesky API fees.
Test results,
artfully arranged
Are you tired of looking at ugly test reports? Testeranto tracks everything in Git and presents test results, static analysis, type errors, screenshots, features, and documentation in a lovely little static site that's easily deployed to GitHub Pages.
'"multlingvulo" means "polyglot" in Esperanto
Testeranto requires Node, and it specializes in Typescript but it also supports Golang and Python. Work has already begun on Rust, Java and Swift implementations. So no matter your native tongue, we have a lingua franca that brings all your tests together.
A new type of BDD
Behavior-Driven Development specifications are written in the language of your choice- no fancy DSL needed. You can edit those specs right in the browser with Github's Web VSCode editor, without ever cloning your project. It's so easy, even a stakeholder can do it.
Your repo made smarter with Aider.
Aider is an AI pair-programming assistant that can read and edit local files. While Testeranto leverages Aider for prompt execution, you can also use it directly to generate tested code right from the command line.
