RAG・知識ベース
llm-graph-builder
Build Neo4j knowledge graphs from unstructured data using LLMs — then chat with your graph.
★ 5.3k スターApache-2.0
ガイドを見る →unstractは、Zipstack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: Zipstack/unstract）。GitHubスターは7,277、フォークは718で、AGPL-3.0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。
Unstract — LLM-driven extraction of unstructured data, built for API deployments and ETL pipeline workflows.
Pick a representative document type and write down the exact fields and JSON shape your downstream system needs.
Clone Zipstack/unstract from GitHub (or use the install path in the run steps below) and skim its README for the parts relevant to your goal.
Clone the repo and start the platform with its Docker Compose setup, keeping the generated encryption key safe — losing it makes existing adapter credentials inaccessible.
Log in locally, build one extraction for your sample document, test it on a few files, then deploy it as an API endpoint.
Extracts structured JSON from unstructured documents using LLMs, aimed at document-AI and IDP workloads.
Extractions are designed to be deployed as APIs, not just run interactively in a notebook.
Fits extraction into ETL pipeline workflows for recurring document processing.
A platform UI for designing and iterating on extraction prompts and schemas.
Run locally with Docker Compose, pull a specific version tag, upgrade in place, or build images locally.
Exposes MCP server capability so agent tooling can call extraction as a tool.
Get the source from the official GitHub repository for unstract.
git clone https://github.com/Zipstack/unstract.git
cd unstractClone the repository and follow the README quickstart to start the platform with Docker.
git clone https://github.com/Zipstack/unstract.git
cd unstract # then start with Docker Compose per the README quickstartVisit the local frontend, log in with the README's default local credentials, and build your first extraction.
# Open http://frontend.unstract.localhost — default local login is in the README; change it after first useZipstack, through the Zipstack/unstract repository, with the product site at unstract.com.
Yes — the project is open source under the AGPL-3.0 license as recorded for this listing. See the license terms for any conditions on redistribution.
Extraction calls LLMs, so you configure your own LLM provider credentials as adapters; Unstract itself is open source under AGPL-3.0, and losing the platform encryption key locks existing adapter credentials.
Template IDP breaks when layouts change and needs per-format rules. Unstract uses LLM extraction defined by prompts and schemas, deployable as APIs and pipelines — more flexible across document variety, with model cost per document instead of template maintenance.
Clone the repository (or use the documented installer), follow the install and configure steps on this page, and run the smallest real task first — log in locally, build one extraction for your sample document, test it on a few files, then deploy it as an API endpoint.
Build Neo4j knowledge graphs from unstructured data using LLMs — then chat with your graph.
All-in-one AI framework for semantic search, LLM orchestration and language model workflows.
AI-native data platform: 60+ connectors, ETL, scheduling, and RAG search.
Lightweight OCR toolkit turning PDFs and images into structured, LLM-ready data in 100+ languages
Ready-to-run templates for RAG, AI pipelines and enterprise search that stay in sync with live data
Libratory, in TypeScript — built for your free book and audiobook laboratory. turn the...
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 — unstract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unstract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。
DeepSearcher — open-source deep research over your private data, with reasoning models and Milvus retrieval.
Vespa — the AI search platform: vector, text and tensor search at production scale.
Utopia — an open-source enterprise world model: a bitemporal knowledge substrate that learns from your documents and gives agents a governed, auditable memory
SQLBot — DataEase's conversational analytics: ask in natural language, get SQL answers via LLMs plus RAG.
Deep Lake — a multimodal datalake and vector store for AI, with version control and training streaming.
Build a ChatGPT over your own data using natural language and LlamaIndex.