TestMu AI adds native API calling to Kane CLI test runner

TestMu AI, the San Francisco and Noida-based quality engineering vendor formerly known as LambdaTest, has added native API calling to its Kane CLI browser-testing tool. The capability, available immediately via npm, allows a test flow to make an API call, store the response by name, and reference that value in subsequent browser steps within the same execution run.

The change addresses a longstanding friction point in end-to-end testing. Until now, teams typically assembled a patchwork of shell scripts and dedicated API clients alongside their browser tests, passing data between tools with glue code that was itself a source of flakiness. By treating an API call as just another natural-language step inside Kane CLI, the vendor removes that seam entirely.

What the capability does

In practice, the feature supports patterns such as creating a database record through an API before asserting its status on screen, fetching an authentication token and injecting it into a browser session, or conditionally branching a test based on whether a live inventory endpoint returns a non-zero stock figure. Stored responses propagate through Kane CLI's nested flow structure: a token or response captured in a parent flow is available at any nesting depth without being re-fetched.

Observability has also been strengthened in the same release. Every API capture is explicitly logged, and branching decisions that depend on API responses are written to the run's execution trail alongside screenshots and step traces, giving engineers a full audit path when a test fails.

Mudit Singh, co-founder and head of growth at TestMu AI, said: "By making API calls a native step inside a natural language flow, the same backend response that creates the state also decides what the UI must display. Assertions can never drift away from the data behind them."

Market context

AI-assisted testing tools are proliferating rapidly. Established players such as Playwright (Microsoft), Cypress and Selenium continue to anchor the open-source end, while a growing cohort of commercial vendors including Mabl, Testim, Applitools and Reflect are embedding AI models to reduce test-authoring overhead and improve self-healing capability. TestMu AI's positioning around natural-language test authorship places it in direct competition with this second cohort, with the additional claim that its agentic architecture can reason across multi-step flows rather than simply record-and-replay interactions.

The move to native API calling within a browser test runner is noteworthy because it tightens the feedback loop between contract testing and UI testing, two disciplines that engineering teams often run in separate toolchains. Vendors that can credibly unify those layers stand to reduce both tooling sprawl and CI pipeline complexity, which are recurring procurement concerns for quality engineering leads at mid-to-large software organisations.

Standards and compliance considerations

Agentic test tooling that calls live backend APIs during a test run carries data-handling implications. Where test environments share infrastructure with staging or production systems, API calls made by automated agents may touch personal or financially sensitive data. Buyers operating under GDPR, SOC 2 or ISO 27001 frameworks will want assurance that Kane CLI's API call logging and response storage do not inadvertently persist regulated data outside approved boundaries. TestMu AI's release does not address this point, and prospective enterprise customers should seek clarification on data-residency and log-retention policies before deploying the feature against non-anonymised environments.

The capability is available now in Kane CLI through the npm registry. TestMu AI did not announce pricing changes or disclose the size of its existing customer base in the release.