Greptile launches codebase call-graph code review with multi-agent investigation

Greptile was founded by Daksh Gupta, Soohoon Choi, and Vaishant Kameswaran — all Georgia Tech CS graduates from the class of 2023 — with the observation that most AI code review tools analysed the diff of a pull request in isolation. Real bugs, however, often lived not in the changed lines themselves but in how the changed code interacted with other parts of the system: a function renamed in one file, still called by its old name in forty others.

Greptile’s response was to build a language-agnostic call graph of the entire repository before any review ran. When a PR arrived, it deployed multiple parallel AI agents that navigated this graph — tracing dependencies, checking git history for related changes, following leads across files — rather than processing the diff sequentially. The result was reviews that caught architectural implications as well as local defects.

Greptile raised a $25M Series A from Benchmark Capital in September 2025, bringing total funding to $30M at a $180M valuation. By 2026 it had onboarded 9,000+ teams including Stripe, Brex, Substack, and Amazon. Greptile v3 introduced the Anthropic Claude Agent SDK for autonomous investigation and long-term memory that learned from team comment patterns to reduce false positive rates over time.