A repository-aware code reviewer with real enterprise features, but its headline accuracy claim is self-published and worth checking against your own codebase rather than taking at face value.
Greptile is an AI agent that reviews pull requests on GitHub and GitLab. Instead of only looking at the lines that changed, it builds a graph index of the whole repository so it can catch issues that span multiple files. It launched in 2023, went through Y Combinator's Winter 2024 batch, and has since raised a $25 million Series A led by Benchmark Capital.
Greptile connects to GitHub Cloud, GitHub Enterprise Server, GitLab Cloud, or GitLab Self-Managed, or runs self-hosted for enterprise customers.
It builds a graph index of files, functions, and dependencies across the repository rather than looking only at the changed lines.
When a PR opens, Greptile posts inline comments, a PR summary, and sequence diagrams, typically within a few minutes, with each comment carrying a confidence score.
Its TREX feature, in public beta, can generate and run unit tests for the changed code in a sandbox, at an added credit cost.
A 'Fix with your Agent' button on each comment sends the file path, line number, and suggested code to Claude Code, Codex, Cursor, Devin, or Conductor.
AI synthesis of external reviews · not on bestaiq
Synthesized from 5 external reviews. Independent signal (Trustpilot / Reddit / verified aggregators) weighted higher than commission-carrying review sites.
Greptile has a free tier for one developer with 50 review credits a month. The Pro tier is $30 per seat per month with 50 credits per seat and $1 for each additional credit. Enterprise pricing is custom and requires contacting sales.
No. It comes from a benchmark that Greptile designed, ran, and scored itself in July 2025. When DeepSource and competitor Augment Code re-scored the same five repositories with different judging rules, they reported a catch rate closer to 45% for Greptile, so treat the 82% figure as a vendor claim rather than an independently confirmed result.
Greptile's terms allow it to use de-identified, aggregated customer data to train its models, but customers can opt out from their account settings. For self-hosted deployments, no customer data is used for machine learning unless the customer explicitly configures it.
Yes. Enterprise customers can deploy Greptile on-prem using Docker Compose or Kubernetes, with a choice of LLM provider including OpenAI, Anthropic, AWS Bedrock, Azure OpenAI, or GCP Vertex AI. Self-hosting requires a separate license.
GitHub Cloud, GitHub Enterprise Server, GitLab Cloud, and GitLab Self-Managed. It does not currently support Bitbucket or Azure DevOps.
Greptile states it is free for qualifying non-commercial projects licensed under MIT or Apache. Some maintainers have reported being billed anyway, so it is worth confirming terms directly with Greptile before relying on the free tier for a production open source project.
Greptile reviews entire pull requests against a graph of the whole repository, which lets it catch issues a diff-only tool would miss, and it offers real enterprise features like self-hosting and a choice of LLM provider. Its central performance claim, an 82% bug catch rate against competitors, is self-published and was scored much lower when independent parties re-ran the same test, so buyers should verify results against their own codebase rather than the vendor's number. The March 2026 move to per-review pricing also makes it worth modeling actual monthly PR volume before committing to Pro.