RAG・知識ベース
graphrag
Microsoft's graph-based RAG: index a corpus into a knowledge graph, then query it with an LLM
★ 36.3k スターMIT
ガイドを見る →code-graph-ragは、vitali87が公開しているオープンソースのRAG・知識ベースプロジェクトです（GitHub: vitali87/code-graph-rag）。GitHubスターは5,238、フォークは699で、MITライセンスの下で公開されています。GitHubでの最終更新は2026年10月8日です。
RAG for your monorepo: query, understand and edit multi-language codebases with a knowledge graph.
Decide what you need: understanding a monorepo, finding dead code, tracing behavior, or safe automated edits.
Install code-graph-rag with the Tree-sitter and semantic extras via uv or pipx.
Start the packaged daemon, then parse your repository into the graph.
Query the codebase in natural language and inspect the grounded, graph-backed answer.
Parses source with Tree-sitter into AST-level graph structure — functions, classes, calls and references — stored in Memgraph.
Asks questions in plain language; an AI model generates Cypher against the graph and answers grounded in the real structure.
Edit code through the agent with AST-based patching and a diff preview before anything changes.
Finds dead code by walking call and reference edges from entry points, and searches/rewrites by AST pattern with ast-grep.
cgr trace merges actually-executed calls (test runs or production eBPF profiles) into the graph, exposing dynamic dispatch static analysis cannot see.
Exposes the graph to MCP-compatible agents, with incremental indexing that keeps the graph current as code changes.
Clone Code-Graph-RAG if you want the source; most users install the tool.
git clone https://github.com/vitali87/code-graph-rag.git
cd code-graph-ragInstall the packaged tool with full Tree-sitter language support and semantic search.
uv tool install "code-graph-rag[treesitter-full,semantic]"Start the packaged Memgraph + Qdrant stack, parse your repo into the graph, and start querying.
cgr daemon up
cgr start --repo-path /path/to/repo --update-graphIt is maintained by vitali87 and contributors, with enterprise services offered around the open-source core.
Yes — the tool is MIT licensed and runs locally; enterprise cloud and on-premise services are optional paid offerings.
Querying uses an AI model to generate Cypher, so you need access to a model provider you configure; the indexing itself is local.
Search tools return chunks; Code-Graph-RAG builds a structural knowledge graph you can traverse — call edges, references, runtime traces — and edit through, at the cost of a heavier indexing stack.
Install with uv, run cgr daemon up, and index your repo with cgr start --repo-path.
Microsoft's graph-based RAG: index a corpus into a knowledge graph, then query it with an LLM
Graph RAG conversational knowledge base built on TiDB Serverless vector storage
Open rag & knowledge for "autonomie" is a highly semantic, responsive, accessibl.
A low-code MCP framework for building complex, innovative RAG pipelines — visibly, not as a black box.
A new state of the art for RAG: an open knowledge base for humans and agents.
Free rag & knowledge tool that 🥥 coco ai server - search, connect, collaborate, ai-pow.
LibreAIカタログに掲載されたRAG・知識ベースプロジェクトの中で最もスターの多いもの — code-graph-ragが合わなかった場合に検討したい、現実的な候補リストです。
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
カタログの中でcode-graph-ragとスター数が近いRAG・知識ベースプロジェクト — 前後の棚もそのまま見て回れます。
Build Neo4j knowledge graphs from unstructured data using LLMs — then chat with your graph.
An agent that finds anything on your computer — federated search over your files, in place.
A low-code MCP framework for building complex, innovative RAG pipelines — visibly, not as a black box.
OpenKB compiles your documents into a persistent, interlinked wiki knowledge base — no vector DB
An OLTP graph database with native vector and full-text search, built in Rust on object storage.
AI-native database for hybrid search across dense, sparse, and tensor vectors plus full text