# Chonkie

A lightweight Python library for splitting text into chunks for retrieval-augmented generation pipelines, with more than thirty integrations and support for 56 languages.

Chonkie is a text-chunking library for retrieval-augmented generation pipelines. It installs with a minimal core plus optional extras for specific chunkers, or a full install that pulls in every dependency, and can be used as a plain library or chained into a pipeline that fetches, chunks, refines and embeds text before it ships to a vector database.

It can also run as a self-hosted REST API server, started from its command line or directly with a server such as uvicorn, so other applications can chunk text over HTTP instead of importing the library. It supports more than thirty integrations with other tools and vector databases and works with 56 languages out of the box.

Website: https://docs.chonkie.ai