Augment Code
Enterprise AI coding platform built for large, complex codebases

Codebase-aware AI reviewer that catches bugs
Greptile is a strong pick when catching real, cross-file bugs matters more than a quiet inbox: independent benchmarks credit it with a high bug-catch rate, but also more false positives than some rivals, so expect to triage. Its 2026 move to per-review usage pricing was divisive and worth modeling against your PR volume before committing. The whole-codebase indexing is its real differentiator. Verify current pricing and the free-tier limits with the vendor.
Greptile is a codebase-aware AI code reviewer that indexes the whole repository to reason about how pull-request changes affect the rest of the code. It scores highly on bug-catch rate in independent tests, with more false positives than quieter rivals. Founded out of YC W2024 and backed by Benchmark, it uses a free tier plus usage-based per-review pricing introduced in 2026.
Greptile is an AI code reviewer built around deep codebase understanding: rather than reviewing a diff in isolation, it indexes the whole repository to reason about how changes interact with the rest of the code. In independent comparisons it has scored notably high on bug-catch rate, at the cost of more false positives than some quieter competitors, reflecting a deliberate tilt toward finding real issues. Founded by Georgia Tech graduates and part of Y Combinator's W2024 cohort, Greptile raised a Series A from Benchmark. In 2026 it shifted to a usage-based pricing model, billing per review beyond a free monthly allowance, and later launched a general free tier with 50 reviews per month for unlimited authors. The per-review model drew some developer pushback as an unusual choice in a category that mostly charges per seat. Greptile suits teams that prioritize catching subtle, cross-file bugs and are willing to triage more flagged items to get them. Teams that want the lowest-noise experience may prefer alternatives, but for aggressive detection with whole-repo context, Greptile is a strong option.
Greptile is a codebase-aware AI code reviewer that indexes whole repositories to catch cross-file bugs, scoring high on bug detection in independent tests at the cost of more false positives. It came out of YC W2024, raised a Series A from Benchmark, and in 2026 moved to usage-based per-review pricing alongside a free tier. It suits teams that prioritize aggressive detection over the quietest experience.
Greptile was founded in 2023 by Georgia Tech graduates Daksh Gupta, Soohoon Choi, and Vaishant Kameswaran, and joined Y Combinator's W2024 cohort.
The company focuses on codebase-aware AI code review, differentiating on whole-repository understanding, and has attracted notable venture backing as demand for automated review grows.
Greptile indexes the entire codebase to review pull requests with full context, flagging bugs and suggesting fixes based on how changes interact across files. It integrates with GitHub, GitLab, and Bitbucket and offers an API.
Its differentiator is depth of context, which drives a high bug-catch rate; the trade-off is more false positives. Pricing combines a free monthly review allowance with usage-based per-review billing.
Engineering teams, especially those with complex, interdependent codebases, that prioritize catching subtle bugs and are willing to triage more flags to get them.
Software developers whose pull requests receive codebase-aware reviews and bug flags.
Engineering managers and platform leads choosing a reviewer tuned for detection.
Senior engineers benchmarking bug-catch rate and false-positive trade-offs.
An engineering team with a complex codebase that values aggressive, context-aware bug detection and can absorb some extra triage and usage-based costs.
Greptile raised a $25M Series A from Benchmark in September 2025, bringing total disclosed funding to about $30M at a reported ~$180M valuation. Verify current figures with the vendor.
It indexes your entire codebase rather than reviewing a diff in isolation, which helps it catch cross-file bugs that diff-only reviewers miss.
It catches more bugs than many rivals but also produces more false positives, so teams should expect some triage in exchange for higher detection.
There is a free tier with 50 reviews per month. Paid usage is largely per-review beyond an allowance, an unusual model that can be cheap at low volume but variable at scale.
Greptile came out of Y Combinator's W2024 batch and raised a Series A led by Benchmark.
No. It is an automated reviewer that surfaces bugs and context, but human judgment is still required before merging.
Side-by-side pages for pricing, features, and best-fit use cases.
Enterprise AI coding platform built for large, complex codebases
Query-aware prompt compression that cuts LLM input tokens by roughly 60% before inference.
Run open LLMs locally with a single command.
AI code quality platform for agentic PR review, testing, and SDLC governance