novetest
latestv0.1.3novetest is the testing intelligence layer for AI coding agents — continuous, local, and reproducible. AI writes code and tests, but native testing engines were built for humans reading a screen. novetest sits on top of those native runners and turns their raw output into structured signal an agent can act on.
One install drops it into any repo. From there a single command runs the suite, normalizes every result, and writes a machine-readable report — closing the gap between the productivity and the trustworthiness of AI-generated code. For a hands-on feel of coverage-guided test design, try the free Test Case Generator in your browser.
What makes novetest, novetest.
An AI-first testing tool.
Structured JSON output built for AI coding agents to read — with a pretty, human-readable format for you whenever you want it.
One workflow, six languages.
The same commands test projects across six ecosystems, each wrapping its native runner. It coexists with Pytest for Python, Jest for JS/TS, JUnit 5 for Java, Go test for Go, dotnet test for C#/.NET, and Cargo test for Rust.
A whole suite of testing tools.
A single testing-intelligence layer that bundles every capability your agent needs into one tool. It contains contract, coverage, memory, regression, localization, replay, and orchestration.
Open source and fully local.
Free to use and Apache-2.0 licensed. No API calls, no remote access — every test runs fast on your own machine.
Runs on every platform.
Fully supported on Windows, Linux, and macOS — wherever your agent and your code already live.
An AI-native TDD ecosystem.
Together with the Novetest product family to come, the foundation of an ecosystem for AI-native test-driven development.
An intelligence layer over your native runners.
Native runners stay at the bottom. novetest layers contract, coverage, memory, regression, localization, replay, and orchestration on top — and exposes the whole stack through one interface. Every engine's commands are covered in the advanced usage guide.
Orchestration
Delivers the entire novetest workflow at once through a single, simple interface.
Regression
Catch regression failures the moment they appear.
Localization
Localize the suspicious root cause behind a failure.
Replay
Detect flaky tests by replaying under controlled conditions.
Memory
Preserves and manages test results across runs — a durable history the agent can reason over.
Coverage
Test coverage and adequacy metrics — a measured answer to "are these tests enough?"
Contract
Unifies the heterogeneous output of each language's native tool into one consistent contract.
Built for two kinds of user.
Developers & managers
Developers who want advanced testing capabilities that go beyond the native tools — layered on top of them, not replacing them.
Managers who want the AI coding agents they run to develop with proper, trustworthy testing in the loop.
Coding & testing agents
Any agent told to run tests after writing code — across any of the supported languages.
Agents that want accurate interpretation, analysis, and next-action recommendations from a tool built with AI as its primary user.