What's new in Tusk

This March, we've rolled out key improvements to make Tusk faster and more powerful for your unit testing needs.

We've integrated SOTA models like Claude 3.7 and Gemini 2.5 into our pipeline. These models allow Tusk to better understand your code and create more relevant, comprehensive tests with less guidance.

Speed was a major focus in this release. Our new parallel test execution capability distributes work across multiple test sandboxes, dramatically reducing latency and getting you results faster than ever before.

We've also given you greater flexibility in accepting Tusk-generated tests. You can now select specific Tusk tests (at a test case or symbol level) to incorporate into your PR/MR, as well as incorporate tests from older commits.

Release notes

Intelligent usage of SOTA models (Claude 3.7, Gemini 2.5)
Improved context gathering for test generation
Ability to load balance across multiple test sandboxes, reduced latency with parallel test execution
Flexibility when incorporating tests - specific tests, older commits
Ability to provide feedback on individual tests
Ability to retry test generation
Show code coverage gains for pytest and Jest
Enhanced test iteration (e.g., learn from iteration history, better determination of required action)
UI/UX improvements for Tusk’s test generation page and PR/MR comments
Routing requests through static IPs for on-prem GitHub/GitLab deployments
Better support for Ruby repos

TRY TUSK NOW

AI test generation with codebase context for quality-obsessed teams.

Cover your blind spots, catch verified critical bugs, merge PRs 25% faster with peace of mind.