RAG и базы знаний
llm-graph-builder
Build Neo4j knowledge graphs from unstructured data using LLMs — then chat with your graph.
★ 5,3 тыс. звёздApache-2.0
Смотреть руководство →rags — проект с открытым исходным кодом из категории «RAG и базы знаний» от run-llama (run-llama/rags на GitHub). У него 6 549 звёзд на GitHub и 654 форка, он распространяется по лицензии MIT и последний раз обновлялся на GitHub 5 апреля 2024.
Build a ChatGPT over your own data using natural language and LlamaIndex.
Decide which documents or data sources you want to chat with and what questions the app should answer.
Clone the repo and install dependencies with Poetry.
Add your OpenAI key (or configure Anthropic, Replicate or a local Hugging Face model) in the app's configuration.
Use the builder in natural language to configure chunking, embeddings and retrieval, then chat with the generated RAG agent.
Describe what you want in plain language and the builder agent assembles the RAG application for you.
Tune the system prompt, top-k retrieval, chunk size, embedding model, LLM and whether to add a summarization tool.
Supports OpenAI, Anthropic and Replicate models, plus local Hugging Face models referenced as local:<model_name>.
Uses OpenAI text-embedding-ada-002 by default and also supports local Hugging Face embedding models.
The generated app runs as a Streamlit UI with a home page, a RAG config page and the generated agent page.
Uses the LlamaIndex framework from run-llama for indexing and querying your data.
Clone RAGs and enter the project directory.
git clone https://github.com/run-llama/rags.git
cd ragsInstall the project and its dev dependencies with Poetry.
poetry install --with devStart the Streamlit interface, then build your RAG agent from the home page.
streamlit run 1_🏠_Home.pyRAGs is a run-llama project — the team behind LlamaIndex.
Yes, the code is MIT licensed. You pay only for any model APIs you choose to call.
By default it uses OpenAI models and embeddings, which need an OpenAI API key; Anthropic and Replicate also need keys, while local Hugging Face models can run without one.
RAGs is a builder on top of LlamaIndex: faster to get a working chat app via natural language, but less flexible than hand-writing your own LlamaIndex pipeline.
Clone the repo, run poetry install --with dev, add your model key, and launch the Streamlit app.
Build Neo4j knowledge graphs from unstructured data using LLMs — then chat with your graph.
SQLBot — DataEase's conversational analytics: ask in natural language, get SQL answers via LLMs plus RAG.
A knowledge-base platform with visual AI workflows — build Q&A systems on your own documents
DeepSearcher — open-source deep research over your private data, with reasoning models and Milvus retrieval.
Polyglot document intelligence in Rust — extract text, tables and structured data from 106 formats.
Unstract — LLM-driven extraction of unstructured data, built for API deployments and ETL pipeline workflows.
Другие проекты из категории «RAG и базы знаний» с наибольшим числом звёзд в каталоге LibreAI — реалистичный список альтернатив, если rags вам не подходит.
Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or…
RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that fuses cutting-edge RAG with Agent capabilities to create a superior context…
Lightweight OCR toolkit turning PDFs and images into structured, LLM-ready data in 100+ languages
MinerU — turns PDFs and Office documents into LLM-ready Markdown and JSON for agents and RAG pipelines, Apache-2.0 plus terms
Stop renting your intelligence. Own it with AnythingLLM. Everything you need for a powerful local-first agent experience
Соседние проекты из категории «RAG и базы знаний» в каталоге, ближайшие по числу звёзд к rags — продолжайте изучать полку по обе стороны от него.
SQLBot — DataEase's conversational analytics: ask in natural language, get SQL answers via LLMs plus RAG.
An OLTP graph database with native vector and full-text search, built in Rust on object storage.
Vespa — the AI search platform: vector, text and tensor search at production scale.
A low-code MCP framework for building complex, innovative RAG pipelines — visibly, not as a black box.
Unstract — LLM-driven extraction of unstructured data, built for API deployments and ETL pipeline workflows.
Build Neo4j knowledge graphs from unstructured data using LLMs — then chat with your graph.
Бесплатно и открыто — забирайте исходный код, ставьте звезду репозиторию и начинайте создавать.