RAG・知識ベース
infinity
AI-native database for hybrid search across dense, sparse, and tensor vectors plus full text
★ 4.7k スターApache-2.0
ガイドを見る →helix-dbは、HelixDB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: HelixDB/helix-db）。GitHubスターは6,117、フォークは375で、Apache-2.0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。
An OLTP graph database with native vector and full-text search, built in Rust on object storage.
Sketch the entities and relationships you need, and note where vector similarity or keyword search must combine with graph traversal.
Install the helix CLI with the official install script.
Run helix init to scaffold helix.toml and an example query, and configure whether local data stays in memory or persists to disk.
Start the local dev instance and run your first query from the CLI or your app's SDK.
Model a labeled property graph, run approximate nearest-neighbor vector search prefiltered by graph traversal, and BM25 keyword search over node and edge properties.
Built for transactional workloads with ACID guarantees; on Helix Cloud a single writer commits every transaction and readers auto-scale.
Durable data lives on object storage rather than a single local disk, separating storage from compute.
Query the graph with a Cypher-inspired language from Rust, TypeScript, Go or Python SDKs.
helix start dev serves a local instance on http://localhost:6969, and an embedded mode runs HelixDB inside your own process without a server.
The README includes a path for letting an AI agent scaffold and build your Helix app for you (helix chef).
Clone HelixDB if you want the source; most users install the CLI in the next step.
git clone https://github.com/HelixDB/helix-db.git
cd helix-dbInstall the helix command-line tool on macOS or Linux.
curl -sSL "https://install.helix-db.com" | bashCreate a project, start the local instance, and run the generated example query.
mkdir my-helix-app && cd my-helix-app
helix init local
helix start dev
helix query dev --file examples/request.jsonHelixDB is maintained by the HelixDB team, with docs at docs.helix-db.com and a managed Helix Cloud offering.
The database is open source under Apache-2.0 and free to self-host; Helix Cloud is the paid managed deployment.
No key is needed for local development. You only need an account if you choose Helix Cloud or call external embedding models.
Neo4j is a mature graph database and you typically add a separate vector store; HelixDB combines graph, vector and full-text search in one Rust engine designed for AI workloads.
Install the CLI, run helix init local, helix start dev, and execute the example query.
AI-native database for hybrid search across dense, sparse, and tensor vectors plus full text
seekdb — the AI-native search database unifying vector, text and structured data for agents.
Graph RAG conversational knowledge base built on TiDB Serverless vector storage
AST-powered code intelligence engine — graph database plus hybrid semantic search across 62 languages, zero LLM dependency.
Vespa — the AI search platform: vector, text and tensor search at production scale.
A blazing-fast vector index in Rust, built on TurboQuant - with Python bindings
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 — helix-db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helix-db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 a ChatGPT over your own data using natural language and LlamaIndex.
A low-code MCP framework for building complex, innovative RAG pipelines — visibly, not as a black box.
SQLBot — DataEase's conversational analytics: ask in natural language, get SQL answers via LLMs plus RAG.
Build Neo4j knowledge graphs from unstructured data using LLMs — then chat with your graph.
Vespa — the AI search platform: vector, text and tensor search at production scale.
RAG for your monorepo: query, understand and edit multi-language codebases with a knowledge graph.