FlashRAG

⚡FlashRAG: A Python Toolkit for Efficient RAG Research (WWW2025 Resource)

What is FlashRAG?

⚡FlashRAG: A Python Toolkit for Efficient RAG Research (WWW2025 Resource)

Overview of FlashRAG

What it is

FlashRAG is a Python toolkit designed for the reproduction and development of Retrieval Augmented Generation (RAG) research. It includes 36 pre-processed benchmark RAG datasets and 23 state-of-the-art RAG algorithms, among which 7 are reasoning-based methods that combine reasoning ability with retrieval. The toolkit provides an easy-to-use UI for configuring and running RAG baselines and evaluation scripts. Its technical paper has been accepted to the Resource Track of the 2025 ACM Web Conference (WWW 2025).

Key features

  • Extensive and customizable framework with components such as retrievers, rerankers, generators, and compressors for flexible pipeline assembly.
  • Comprehensive benchmark datasets: 36 pre-processed RAG datasets for testing and validation.
  • Pre-implemented advanced RAG algorithms: 23 algorithms with reported results, including 7 reasoning-based methods.
  • Efficient preprocessing stage with scripts for corpus processing, retrieval index building, and pre-retrieval of documents.
  • Optimized execution using tools like vLLM, FastChat for LLM inference acceleration, and Faiss for vector index management.
  • Easy-to-use UI for configuring and experiencing supported RAG methods and running evaluation scripts visually.

Use cases

FlashRAG is suitable for researchers and developers who need to reproduce existing state-of-the-art RAG works, implement custom RAG processes and components, or evaluate RAG models on benchmark datasets. It supports tasks such as multi-hop reasoning, multimodal RAG, and domain-specific RAG testing.

Getting started

To install FlashRAG with pip:bashpip install flashrag-dev --preOr clone from GitHub and install (requires Python 3.10+):bashgit clone https://github.com/RUC-NLPIR/FlashRAG.gitcd FlashRAGpip install -e .For optional dependencies like vllm, sentence-transformers, or pyserini, see the repository README.

License and pricing

The license is MIT. The project is open source.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/ruc-nlpirflashrag.svg)](https://yunodb.com/tools/ruc-nlpirflashrag?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/ruc-nlpirflashrag?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/ruc-nlpirflashrag.svg" alt="Featured on YunoDB" height="20"></a>