QAxolotl
by QAxolotl
AI writes code. QAxolotl breaks it — before your users do.
Demo Video
About This Project
QAxolotl is an AI QA Agent that lives inside VS Code and tests AI-generated code the way a real human would — by opening a real browser, clicking real buttons, and watching what actually happens.
In the age of AI-generated code, developers ship faster than ever — but nobody's watching whether the code actually works. Unit tests check logic. Linters check syntax. But does the login button fire? Does the error toast appear? Does the checkout flow survive edge cases? Nobody knows until production breaks.
QAxolotl closes that gap.
Powered by MiniMax-M2.5 as its core reasoning engine, QAxolotl performs intelligent code analysis — reading git diffs, classifying PR types, identifying critical user flows, and generating executable test plans — all without human intervention. MiniMax-M2.5 drives every layer of Axolotl's intelligence: from understanding what changed in a pull request, to deciding what to test, to interpreting what went wrong when a test fails. Its 192K-token context window lets QAxolotl ingest entire codebases and reason across files with surgical precision.
For voice-first developer workflows, QAxolotl integrates Speechmatics real-time speech-to-text, enabling developers to dictate test instructions, describe expected behaviors, and narrate bug reports hands-free — in 50+ languages. Using Speechmatics' enhanced-quality WebSocket API, voice input is transcribed live and fed directly into QAxolotl's QA pipeline, making testing as natural as talking to a colleague.
How it works:
You open a PR. QAxolotl reads the diff and auto-classifies what changed. It injects diagnostic logging into critical code paths. It generates a battle-tested test plan — covering happy paths, edge cases, and failure modes. It launches a real browser, executes each test step, captures screenshots, records console logs, and collects hard evidence. It delivers a verdict: MERGEABLE or NOT_MERGEABLE — with receipts. QAxolotl isn't a test framework. It isn't a CI pipeline. It's your AI QA teammate that reviews every PR with the rigor of a senior engineer and the speed of a machine.
Built with:
MiniMax-M2.5 — Core AI engine for code analysis, test generation, and intelligent reasoning across the entire QA workflow Speechmatics — Real-time multilingual voice-to-text for hands-free test authoring and developer interaction VS Code Extension API — Deep IDE integration for seamless developer experience
Tracks
Built With
Repository
Submitted February 21, 2026 at 4:12 PM