onyx-dot-app/onyx
onyx-dot-app
28.6k
Open Source AI Platform - AI Chat with advanced features that works with every LLM
3.8k|Python
NOASSERTIONenterprise-searchragai-chat
Self-hosted RAG
Tools for building private knowledge bases, document chat, and retrieval-backed AI apps.
Use this page when your AI system needs to search private documents, connect to internal knowledge, or ground answers in your own data.
Apps and building blocks
The list includes ready-to-use document assistants as well as frameworks and engines for custom RAG pipelines.
Private data focus
These projects are useful when retrieval quality, connectors, indexing, and deployment control matter.
Why it works
Document chat apps
AnythingLLM, Onyx, DocsGPT, Kotaemon, Khoj, Verba, and RAG Web UI are closer to ready-to-use private knowledge assistants.
RAG frameworks and engines
LlamaIndex, Haystack, R2R, RAGFlow, txtai, llmware, RAG-Anything, GraphRAG, and LightRAG fit teams building or customizing the retrieval pipeline.
Visual and advanced RAG builders
Dify, Flowise, Langflow, GraphRAG, and LightRAG are useful when teams need workflows, agents, or graph-based retrieval instead of a simple document-chat app.
Curated repositories
Open source RAG and knowledge tools
Mintplex-Labs/anything-llm
Mintplex-Labs
59.1k
The all-in-one AI productivity accelerator. On device and privacy first with no annoying setup or configuration.
6.4k|JavaScript
MITraglmstudiolocalai
arc53/DocsGPT
arc53
17.9k
Private AI platform for agents, assistants and enterprise search. Built-in Agent Builder, Deep research, Document analysis, Multi-model support, and API connectivity for agents.
2k|Python
MITaipythonnatural-language-processing
Cinnamon/kotaemon
Cinnamon
25.3k
An open-source RAG-based tool for chatting with your documents.
2.1k|Python
Apache-2.0chatbotllmsopen-source
run-llama/llama_index
run-llama
49k
LlamaIndex is the leading document agent and OCR platform
7.3k|Python
MITagentsapplicationdata
deepset-ai/haystack
deepset-ai
25k
Open-source AI orchestration framework for building context-engineered, production-ready LLM applications. Design modular pipelines and agent workflows with explicit control over retrieval, routing, memory, and generation. Built for scalable agents, RAG, multimodal applications, semantic search, and conversational systems.
2.7k|MDX
Apache-2.0nlpquestion-answeringpytorch
SciPhi-AI/R2R
SciPhi-AI
7.8k
SoTA production-ready AI retrieval system. Agentic Retrieval-Augmented Generation (RAG) with a RESTful API.
627|Python
MITartificial-intelligencelarge-language-modelsretrieval-augmented-generation
infiniflow/ragflow
infiniflow
79.1k
RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that fuses cutting-edge RAG with Agent capabilities to create a superior context layer for LLMs
9k|Python
Apache-2.0llm-app
weaviate/Verba
weaviate
7.7k
Retrieval Augmented Generation (RAG) chatbot powered by Weaviate
849|Python
BSD-3-ClauseRetrieval-Augmented Generation (RAG) & Knowledge
rag-web-ui/rag-web-ui
rag-web-ui
3k
RAG Web UI is an intelligent dialogue system based on RAG (Retrieval-Augmented Generation) technology.
335|TypeScript
Apache-2.0ailangchainrag
microsoft/graphrag
microsoft
32.5k
A modular graph-based Retrieval-Augmented Generation (RAG) system
3.4k|Python
MITgraphragragllm
HKUDS/LightRAG
HKUDS
34.4k
[EMNLP2025] "LightRAG: Simple and Fast Retrieval-Augmented Generation"
4.9k|Python
MITknowledge-graphlarge-language-modelsretrieval-augmented-generation
khoj-ai/khoj
khoj-ai
34.3k
Your AI second brain. Self-hostable. Get answers from the web or your docs. Build custom agents, schedule automations, do deep research. Turn any online or local LLM into your personal, autonomous AI (gpt, claude, gemini, llama, qwen, mistral). Get started - free.
2.2k|Python
AGPL-3.0semantic-searchemacsobsidian-md
langgenius/dify
langgenius
139.3k
Production-ready platform for agentic workflow development.
21.8k|TypeScript
NOASSERTIONaigptllm
FlowiseAI/Flowise
FlowiseAI
52.3k
Build AI Agents, Visually
24.2k|TypeScript
NOASSERTIONartificial-intelligencechatgptlarge-language-models
langflow-ai/langflow
langflow-ai
147.4k
Langflow is a powerful tool for building and deploying AI-powered agents and workflows.
8.9k|Python
MITreact-flowchatgptlarge-language-models
neuml/txtai
neuml
12.4k
💡 All-in-one AI framework for semantic search, LLM orchestration and language model workflows
803|Python
Apache-2.0pythonsearchnlp
llmware-ai/llmware
llmware-ai
14.9k
Unified framework for building enterprise RAG pipelines with small, specialized models
2.9k|Python
Apache-2.0parsingretrieval-augmented-generationagents
HKUDS/RAG-Anything
HKUDS
18.9k
"RAG-Anything: All-in-One RAG Framework"
2.2k|Python
MITmulti-modal-ragretrieval-augmented-generation
Selection guide
How to choose a self-hosted RAG tool
Separate ready-to-use knowledge-base apps from developer frameworks and retrieval infrastructure. The right choice depends on whether users need a product UI or engineers need a pipeline.
For private knowledge assistants
Onyx, AnythingLLM, DocsGPT, Kotaemon, Khoj, Verba, and RAG Web UI are closer to end-user document chat products.
For custom RAG pipelines
LlamaIndex, Haystack, R2R, RAGFlow, txtai, llmware, RAG-Anything, GraphRAG, and LightRAG are better framed as frameworks, engines, or research-oriented building blocks.
For visual RAG builders
Dify, Flowise, and Langflow fit teams that want to assemble knowledge-base apps, agents, and retrieval flows without starting from raw framework code.
Production fit
Apps, frameworks, and research projects are not the same
A rankable RAG page should make the distinction clear. Some projects are production-facing apps; others are libraries, methods, or experimental memory systems.
Strongest app fit
Onyx, AnythingLLM, DocsGPT, Kotaemon, and Khoj are the best matches for private knowledge-base and document-chat intent.
Adjacent but useful
Dify, Flowise, Langflow, GraphRAG, and LightRAG are valuable, but should not be sold as simple turnkey document-chat apps.
Retrieval stack
Do you need a vector database?
Many RAG systems use vector search, but the vector database is only one layer. Ingestion, chunking, metadata, reranking, citations, access control, and evaluation can matter more than the store alone.
Use vector stores as infrastructure
Qdrant, Chroma, Weaviate, Milvus, LanceDB, and pgvector support RAG, but they are not complete RAG products by themselves.
Suggested additions
Strong candidates not yet in the registry
PrivateGPT
zylon-ai/private-gpt
A private document Q&A project focused on local/offline RAG workflows. Strong fit for this page once it is present in the upstream registry.
View repositoryFastGPT
labring/FastGPT
A knowledge-base platform with RAG retrieval, data processing, and visual AI workflows. Strong product fit, with licensing details to review before commercial use.
View repositoryMaxKB
1Panel-dev/MaxKB
A self-hostable knowledge-base and agent platform built around RAG pipelines, document ingestion, and enterprise-style assistant workflows.
View repositoryQuivr
QuivrHQ/quivr
A self-hostable second-brain style RAG project. Relevant to private knowledge-base intent, though its current positioning is broader than simple document chat.
View repositoryDB-GPT
eosphoros-ai/DB-GPT
A self-hosted AI data assistant with private knowledge and RAG capabilities. Better for data-heavy assistant workflows than simple file-based document chat.
View repositorylocalGPT
PromtEngineer/localGPT
A local document-chat project focused on private, on-device Q&A. Useful candidate, but narrower and less platform-like than the main curated options.
View repositorySurfSense
MODSetter/SurfSense
A privacy-focused NotebookLM-style project for teams. Promising fit for private knowledge workflows, but newer and less established than the main list.
View repositoryRelated pages
Keep browsing
Self-hosted ChatGPT alternatives
Chat interfaces and assistant apps you can run with local models, private endpoints, or your own hosted providers.
Local model runtimes and inference servers
Inference servers and local runtimes for serving models on your machine, server, or private cloud.
Vector databases and retrieval storage
Databases and search layers for embeddings, metadata filtering, persistence, and semantic retrieval.
Agents, workflows, and app builders
Agent frameworks, workflow engines, and app builders for repeatable AI-powered processes.
AI developer tools
Coding assistants and repo-aware tools that can run locally or inside private development environments.
FAQ
Questions answered
What is a self-hosted RAG tool?
It is an open source project that helps language model apps retrieve private or custom knowledge while running in your own environment as an app, framework, service, or retrieval database.
Do I need a vector database for RAG?
Often, but not always. Many RAG systems use vector search, but smaller or specialized deployments may use hybrid search, full-text search, reranking, or framework-managed indexes.
Should document parsers and crawlers be listed here?
Only when they are part of a broader RAG product. Tools that only parse PDFs, crawl websites, or convert files are useful inputs, but they are not enough for this page on their own.
What is the difference between a RAG app and a RAG framework?
A RAG app gives users a product interface for document search or chat. A RAG framework gives developers building blocks for ingestion, retrieval, reranking, evaluation, or custom pipelines.
Which self-hosted RAG tools are best for teams?
Onyx and AnythingLLM are the strongest fits for team-oriented private knowledge bases. DocsGPT, Khoj, RAGFlow, R2R, Dify, Flowise, Langflow, and Kotaemon can also fit depending on whether the team needs a product UI, API-first retrieval, visual workflows, or deeper document processing.