Overview

kit: Code Intelligence Toolkit

A modular, production-grade toolkit for codebase mapping, symbol extraction, code search, and LLM-powered developer workflows. Supports multi-language codebases via tree-sitter.

kit features a “mid-level API” to build your own custom tools, applications, agents, and workflows: easily build code review bots, semantic code search, documentation generators, and more.

kit is free and open source with a permissive MIT license. Check it out on GitHub.

Installation

Install from PyPI

# Basic installation, it's currently `cased-kit` on pypi
pip install cased-kit

# With OpenAI support
pip install cased-kit[openai]

# With all features (OpenAI, Anthropic, Google, vector search, API)
pip install cased-kit[all]

Install from Source

git clone https://github.com/cased/kit.git
cd kit
uv venv .venv
source .venv/bin/activate
uv pip install -e .

Why Use kit?

kit helps with:

Unifying Code Access: Provides a single, consistent Repository object to interact with files, symbols, and search across diverse codebases, regardless of language.
Deep Code Understanding: Leverages tree-sitter for accurate, language-specific parsing, enabling reliable symbol extraction and structural analysis across an entire codebase.
Bridging Code and LLMs: Offers tools specifically designed to chunk code effectively and retrieve relevant context for large language models, powering smarter AI developer tools.

Core Philosophy

kit aims to be a toolkit for building applications, agents, and workflows. It handles the low-level parsing and indexing complexity, and allows you to adapt these components to your specific needs.

We believe the building blocks for code intelligence and LLM workflows for developer tools should be free and open source, so you can build amazing products and experiences.

Where to Go Next

Dive into the API: Explore the Core Concepts to understand the Repository object and its capabilities.
Build Something: Follow the Tutorials for step-by-step guides on creating practical tools.