yichuan-w/LEANN — [MLsys2026]: RAG on Everything with LEANN.

Description and use

LEANN is an open‑source framework that enables privacy‑preserving Retrieval‑Augmented Generation (RAG) on any kind of data, delivering up to 97 % storage reduction while maintaining fast and accurate inference on a personal device. It combines efficient data orchestration, lightweight embedding storage, and on‑device language models to provide a fully offline, 100 % private RAG solution for diverse applications.

Problem

yichuan-w/LEANN helps turn isolated prompts and tools into repeatable agent workflows.

Outcome

Medium: useful for prototypes or internal workflows, with dependency and maintenance checks before production.

Workflow

• coordinate multi-agent workflows
• add tool-use pipelines
• standardize agent memory

Evidence

[MLsys2026]: RAG on Everything with LEANN. Enjoy 97% storage savings while running a fast, accurate, and 100% private RAG application on your personal device.

Source