Codebase graph indexing
Greptile builds a graph of your repository so reviews can use file, function, and dependency context instead of only reading the diff.
Greptile is an AI code review tool that analyzes pull requests with repository context to catch bugs, logic issues, and style problems before merge.
Greptile is an AI code review tool that analyzes pull requests with repository-level context. It builds a graph index of the codebase, then uses multiple agents to review changes, look beyond the diff, and surface bugs, style issues, and logic problems before merge.
The product is positioned for teams that want automated PR review with customization for their own standards. Its pages describe learning from prior comments, custom rules and context, and workflows that connect review feedback back into tools such as Claude Code, Cursor, Codex, Devin, and MCP-based setups.
Greptile builds a graph of your repository so reviews can use file, function, and dependency context instead of only reading the diff.
Multiple agents review each pull request in parallel to assess the broader impact of changes and flag issues that humans may miss.
The product learns from engineers’ comments over time and adapts reviews to the team’s coding standards and feedback patterns.
Teams can define plain-English rules, attach repo-specific context, or point Greptile at style guides and scoped rule files to shape review behavior.
Review output can include confidence scores, diagrams, PR summaries, and test generation for new or changed code.
Greptile can be configured with severity thresholds, collapsed sections, and review triggers so teams control when and how reviews run.
Use Greptile to review incoming pull requests with repository context, so issues that depend on surrounding files or dependencies are more likely to be caught before merge.
Use custom rules, style guides, and scoped rule files to enforce team-specific conventions across a repository or subset of files.
Use the review summary, confidence score, diagrams, and follow-up chat to understand complex changes quickly during code review.
Use the generated feedback and one-click fixes in Claude Code, Cursor, Codex, or Devin to move from review comments to implementation faster.
Use self-hosting, SSO/SAML, and enterprise support when the review workflow must fit an internal infrastructure or security process.
Greptile reviews pull requests by indexing a codebase, analyzing changes in context, and surfacing issues it finds before merge. The source describes it as an AI code review agent that can also suggest fixes and answer questions in the PR.
The pricing page shows a Pro plan at $30 per seat per month with a 14-day free trial and an Enterprise plan with custom pricing. The site also offers free access for qualified open source projects and a startup discount for eligible pre-Series A teams.
Yes. The pricing page says Enterprise customers can self-host in their own infrastructure, and the FAQ says Greptile can be self-hosted in an AWS environment with your own LLM providers.
Greptile supports GitHub and GitLab, and the site also describes connecting it with external apps and AI coding tools such as Claude Code, Cursor, Codex, Devin, and MCP-based workflows.
The FAQ states that Greptile fully supports Python, JavaScript, TypeScript, Go, Elixir, Java, C, C++, C#, Swift, PHP, and Rust, with most other languages supported at slightly lower response quality.