Hero Light

Greptile is a platform that can search and understand complex codebases in natural language.

There are broadly two ways to use Greptile -

LLM-powered codebase Q&A

Chat with any codebase using our web app, our VS Code plugin or in your terminal using our CLI, available via npm i -g greptile.

Internal AI Developer tools

Build custom internal AI dev tools such as PR review bots, Slackbots, documentation updaters and more, with the power of full codebase context. See examples

Regardless of how you use Greptile, the overall flow looks like this:

  1. Obtain API key and GitHub token: one-time, takes 5 minutes. API support for GitLab and Azure Devops is coming soon.
  2. Submit repository for indexing: one-time, takes 5-50 minutes depending on codebase size, auto-updates index on every push.
  3. Query in natural language: answers with full-codebase context in <3 seconds, responses can be streamed.

​
Motivation

We believe that full codebase-context is the core problem in building AI developer tools. With reliable, full codebase context, developers can build unlimited tooling, from PR review bots, to JIRA ticket annotators, to Slackbots to answer questions about the source code, and much, much more.

Our mission at Greptile is to provide developers with that infrastructure as an easy-to-use API.

​
What will you build?

Here are some examples of internal tools you can build for your team with Greptile’s API:

Codebase Expert Slackbot

Build an expert on your codebase that answers questions in your Slack!

PR Review Bot

Build a bot that writes thoughtful, codebase context-aware comments on your team’s PRs.

Automatic Docs Updater

Build a bot that writes a PR to update your docs every time you push new code, so your docs are always up-to-date.

Ticket Enricher

Automatically add context and instructions to new tickets made in your task manager - Linear, Jira, etc.

…the possibilities are endless!

Quickstart