Chonkie

Chonkie is an ingestion library that chunks, embeds, and stores data for RAG pipelines.

Chonkie is an ingestion library designed to facilitate the creation of RAG pipelines. It functions by allowing developers to chunk, embed, and store data through a unified interface. Users install the library, import it, and perform data processing tasks directly. The library supports various tokenizers, chunkers, embedding models, and vector databases. It is intended for developers building RAG systems who require a lightweight alternative to larger software suites. The tool also offers AI agent skills for integration with platforms like Claude Code, Cursor, and Copilot.

Chonkie at a glance

Best for: Software engineer
Runs on: Python, JavaScript, API
Pricing: Open source
Released: March 2025
Category: AI coding tool
Website: docs.chonkie.ai

Key features

Clean, chunk, embed, refine, and store data
Support for various tokenizers and chunkers
Support for multiple embedding models and vector databases
Python and JavaScript support
API access
AI agent skills for Claude Code, Cursor, and Copilot

Use it for

Building RAG pipeline ingestion logic
Preprocessing and embedding data for vector search
Integrating data ingestion into AI agents

Alternatives to Chonkie

claude-mem — It shares persistent, searchable agent memory across sessions and devices.
jevgrep — Reduces coding agent costs by providing targeted, relevant repository context.
Code Graph RAG — It builds knowledge graphs from codebases using Tree-sitter parsing.
LangChain4j — Integrates LLMs and vector stores directly into Java applications.
Shimmy — Runs local AI models with zero Python or C++ dependencies.
zerostack — Minimal memory footprint and high CPU efficiency compared to JS-based agents.
Engram — Single Go binary with zero dependencies and built-in full-text search.
Graphite — Accelerates review cycles with AI code reviews and stacked pull requests.
AI SDK — Switch providers easily and stream real-time responses without custom parsing.
iii — Adds services as workers in any language without writing glue code.
DataChain — Processes files in S3, GCS, and Azure at warehouse speed.
Juggler — A native Go server with CRDT-based live conversations and recursive branching threads.

All Chonkie alternatives