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 →vespaは、vespa-engine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: vespa-engine/vespa）。GitHubスターは7,121、フォークは745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。
Vespa — the AI search platform: vector, text and tensor search at production scale.
Decide cloud versus self-hosted, and design your first application package — document schema, indexing and a ranking profile — before writing data.
Clone vespa-engine/vespa from GitHub (or use the install path in the run steps below) and skim its README for the parts relevant to your goal.
Start from the official getting-started guide: deploy an application package to Vespa Cloud or to your own Docker-based instance.
Deploy the sample application, feed a few documents, and run a query with a vector-plus-text ranking profile to see hybrid retrieval work.
Vector search, classic text search and structured filtering are evaluated together in a single query path.
Ranking uses tensor computation, so ML model inference happens inside the search engine at query time.
Designed for large corpora and high query loads with horizontal scaling and multinode redundancy.
Serving and recommendation are first-class uses alongside search, sharing the same application model.
Run applications on Vespa Cloud (with a free trial) or operate your own Vespa instance from the same open-source code.
A companion sample-apps collection provides fully functional, production-ready starting points.
Get the source from the official GitHub repository for vespa.
git clone https://github.com/vespa-engine/vespa.git
cd vespaCreate and deploy your first Vespa application package to Vespa Cloud or your own instance, following the official guide the README links.
# Follow https://docs.vespa.ai/en/getting-started.html to deploy a sample app — see READMEFor Vespa itself, set up the documented dev environment (including JDK 17) and build the Java modules.
brew install jenv mvnvm openjdk@17 # macOS dev setup example from the README; see its Building sectionvespa-engine (the Vespa team, originally from Yahoo), through the vespa-engine/vespa repository, with the service at vespa.ai.
Yes — the project is open source under the Apache-2.0 license as recorded for this listing. See the license terms for any conditions on redistribution.
No API key is needed to self-host the Apache-2.0 engine. Vespa Cloud is a paid managed service (with a free trial) if you prefer not to operate it.
A vector database does nearest-neighbour lookup. Vespa is a full search platform — text, vector, structured data, tensor ranking and model serving in one distributed system — more powerful for production search, and correspondingly more system to learn and run.
Clone the repository (or use the documented installer), follow the install and configure steps on this page, and run the smallest real task first — deploy the sample application, feed a few documents, and run a query with a vector-plus-text ranking profile to see hybrid retrieval work.
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.
An OLTP graph database with native vector and full-text search, built in Rust on object storage.
Vector similarity search inside Postgres - store embeddings next to your data
Ragent (容纳智能体): a production-grade Java Agentic RAG platform with hybrid retrieval and an admin console
All-in-one API platform for search, recommendations, RAG and analytics — self-hostable
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 — vespa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vespa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。
Unstract — LLM-driven extraction of unstructured data, built for API deployments and ETL pipeline workflows.
SQLBot — DataEase's conversational analytics: ask in natural language, get SQL answers via LLMs plus RAG.
DeepSearcher — open-source deep research over your private data, with reasoning models and Milvus retrieval.
Build a ChatGPT over your own data using natural language and LlamaIndex.
Utopia — an open-source enterprise world model: a bitemporal knowledge substrate that learns from your documents and gives agents a governed, auditable memory
An OLTP graph database with native vector and full-text search, built in Rust on object storage.