Tusk takes a fundamentally different approach to test generation by using real production traffic as the source of truth rather than relying on developers to imagine test scenarios. The Drift SDK records API calls in production, capturing request-response pairs across thousands of real user interactions. AI then filters and transforms these traces into high-quality unit and API test cases that cover edge cases developers would never think to test manually.
The platform runs generated tests in ephemeral sandboxes and self-iterates when tests fail, eliminating the back-and-forth typically required with AI coding assistants. Tests are automatically matched to relevant code changes in pull requests and run in CI, with 69% of Tusk-generated tests being incorporated into PRs. The system also ingests business context from Jira and Linear to generate tests aligned with team practices and requirements.
Tusk helped Promptfoo add over 2,000 tests and catch 56 bugs across their codebase while automatically maintaining the test suite to remove flaky tests. The platform offers a 14-day free trial with paid plans for teams. PII redaction rules and domain filtering ensure sensitive production data never leaks into test fixtures. Node.js and Python are supported with the open-source SDK and CLI.