Backed by Y Combinator
Create self-healing, end-to-end tests without leaving your coding agent. Built by ex-Stripe engineers.
Trusted by ENGINEERING TEAMS AT
Best-in-class testing infrastructure
Everything you need to make sure your product never breaks.
Self-healing
Made for developers
Reproducible results
Full coverage
full coverage
Test every surface
Lark's agents can test UIs, APIs, SDKs, async workflows, and more.
FAQ
How do I integrate Lark with my CI/CD pipeline?
We provide native integrations with GitHub Actions, GitLab CI, and other popular CI tools.
What types of applications can Lark test?
Tests can be written for dashboards, API endpoints, and more. You can see some sample tests at https://tests.getlark.ai/a/polar/
How is Lark different from something like Playwright?
Playwright tests are hard to write and maintain because it's code. Lark tests are trivial to write because it's natural language. Playwright tests can be brittle and break easily if the product changes. Lark tests are more robust and can handle changes to the product. Playwright tests are for frontend only, Lark tests can be for frontend, backend APIs, and more.
What happens when a test fails in production?
You'll receive instant alerts via Slack, email, or PagerDuty with detailed logs and screenshots to help you quickly identify and resolve the issue. With our CI integration you can also prevent PRs from merging if tests fail.