Gitar, a two-year-old startup founded by Intel Labs and Google veteran Ali-Reza Adl-Tabatabai, came out of stealth Wednesday with a $9 million seed round led by Venrock. The pitch: AI-generated code is flooding enterprise codebases faster than human reviewers can handle, so Gitar deploys its own AI agents to do the reviewing, testing, and CI triage instead.
What's new
The platform sells subscription access to a suite of code-quality agents that handle code reviews, continuous integration diagnostics, and security checks. Engineering teams can also build their own custom agents on top of the platform. Sierra Ventures participated in the round alongside lead investor Venrock. Adl-Tabatabai frames the core product around a simple split: generation produces code, validation makes it trustworthy — and Gitar owns the validation layer end to end.
Why it matters
The "vibe coding" wave has a well-documented downside: AI-written code ships bugs, introduces security issues, and still requires senior engineers to review before it goes to production. That bottleneck is getting worse as AI coding tools become standard. Gitar's bet is that the review process itself can be largely automated, with humans looped in only for exceptions rather than as the default gate. It's a direct challenge to the assumption that meaningful code review requires a human in the loop.
What to watch
Adl-Tabatabai's stated endgame is minimal human code review — Gitar's agents clearing code for production with humans stepping in only when something flags. That's a significant trust claim for enterprise security teams who are already nervous about AI-generated code in the first place. Whether customers actually hand over that much autonomy to an agent pipeline — and how Gitar handles liability when the agent gets it wrong — will be the real test.