TestMu AI, the company formerly known as LambdaTest, released Test.md on May 14: a testing framework built into its Kane CLI that replaces traditional selectors, framework-specific configurations, and static test scripts with markdown-structured files designed to be readable and writable by both humans and AI agents.
The launch targets a specific problem in the agentic development stack. AI agents write code. AI agents review code. But testing infrastructure has remained largely unchanged, built around assumptions of human authorship: CSS selectors, XPath queries, framework-specific syntax that agents struggle to generate reliably and humans struggle to maintain at scale.
The Framework
Test.md uses markdown as its native format. Each test is defined through step-level headings with plain English objectives that function simultaneously as documentation and executable test cases. “AI has fundamentally changed how software is built, but testing workflows have not evolved at the same pace,” Asad Khan, CEO and co-founder at TestMu AI, told GlobeNewswire. “Test.md creates a shared, durable test format that both humans and AI agents can read, write, and execute, without the overhead of traditional frameworks.”
The framework introduces a replay-first execution model. Tests are authored once and replayed deterministically, with Kane CLI deciding when to reuse recorded steps and when to re-author them based on changes in the application flow. This converts one-time validation into continuous regression coverage without proportional increases in effort.
Key technical features include modular test composition through @import for reusable common flows (login, setup, teardown), configuration embedded via front matter for environment-specific variables, and native CI/CD integration with headless, parallel, and agent-mode execution. Each run generates a result artifact with step-level status, execution traces, and shareable evidence.
Context
LambdaTest rebranded to TestMu AI on January 12, 2026, signaling its shift from a browser and device testing platform to what it calls the “world’s first full-stack Agentic AI Quality Engineering platform,” according to the company’s rebrand announcement. The company was recognized in Forrester’s Wave for Autonomous Testing Platforms in Q4 2025, per NewsBreak.
Test.md is the first major product release under the new brand and represents the company’s bet that testing, not code generation, is the next critical infrastructure gap in the agentic development stack.
The Infrastructure Layer Nobody Built
The agent-native development workflow has coding agents, review agents, and deployment automation. What it has lacked is a testing format that agents can produce without human intervention and that humans can audit without deciphering agent-generated test scripts. Test.md’s markdown approach is an attempt to solve that by making the test artifact the shared contract between human and agent.
Whether the markdown-as-test-format approach scales to complex integration and end-to-end testing remains to be proven in production. But the problem Test.md targets is real: as agent-generated code volume increases, the verification bottleneck moves from “can we write tests?” to “can agents and humans share a testing language?” Test.md is available now as part of Kane CLI across local, cloud, and CI/CD environments.