All-in-one AI app for chatting with your documents and running AI agents, with local-first setup, multi-user support and many LLM providers.
deeplethe/utopia
A self-hosted Rust knowledge base that builds a bitemporal knowledge graph from documents, with RAG search, ontology-driven reasoning and MCP access for agents.
- Звёзды
- 8,445
- Форки
- 1,187
- Открытые issues
- 64
- Последний коммит
- 4 часа назад
Что это?
Utopia is a self-hosted knowledge base that ingests documents, extracts entities and facts into a knowledge graph, and answers questions with cited evidence. Every fact records when it held in the world and when the system learned it, so a correction closes the old version and links the new one instead of overwriting it. Ontologies can be chosen at knowledge base creation, and axioms such as transitivity or inverses can derive new facts when enabled.
The system ships as one Rust binary backed by PostgreSQL with pgvector, with full-text search embedded via Tantivy and a web console for the graph, ontology and review queue. It works with any OpenAI-compatible model endpoint, so it can run without external network access. An MCP server exposes each knowledge base to agents such as Claude Desktop or Cursor with scoped permissions. The project is at v0.1, and its schema migrations only roll forward.
Ключевые возможности
- 01Ingests PDF, Office files, spreadsheets, Markdown, HTML and plain text, with OCR and transcription via external services
- 02Bitemporal graph where corrections close old facts and link to replacements
- 03Hybrid search combining Tantivy full-text and pgvector vectors with RRF fusion
- 04Ontology packs and optional axiom-based reasoning with a human review queue
- 05MCP server for connecting agents with per-knowledge-base permissions
- 06Works with any OpenAI-compatible model endpoint, including local models via Ollama or vLLM
Сценарии использования
- Tracking contract terms and amendments so that answers reflect the terms in force on a given date
- Keeping records of who held a role when, with corrections preserved in history
- Letting coding or operations agents query a company knowledge base over MCP
- Running a document knowledge base on infrastructure without external network access
Темы
Частые вопросы
- Что такое utopia?
- A self-hosted Rust knowledge base that builds a bitemporal knowledge graph from documents, with RAG search, ontology-driven reasoning and MCP access for agents. Utopia is a self-hosted knowledge base that ingests documents, extracts entities and facts into a knowledge graph, and answers questions with cited evidence. Every fact records when it held in the world and when the system learned it, so a correction closes the old version and links the new one instead of overwriting it. Ontologies can be chosen at knowledge base creation, and axioms such as transitivity or inverses can derive new facts when enabled.
- utopia можно использовать бесплатно? Какая у него лицензия?
- utopia распространяется с открытым исходным кодом под лицензией Apache-2.0. Перед использованием в коммерческих продуктах проверьте условия лицензии.
- utopia всё ещё поддерживается?
- Активно поддерживается. Последний коммит в utopia — 19 окт. 2026 г..
- Для чего можно использовать utopia?
- Tracking contract terms and amendments so that answers reflect the terms in force on a given date Keeping records of who held a role when, with corrections preserved in history Letting coding or operations agents query a company knowledge base over MCP Running a document knowledge base on infrastructure without external network access
- Где скачать utopia?
- Исходный код и инструкции по установке — на GitHub: https://github.com/deeplethe/utopia
Описание создано ИИ на основе README проекта. Статистика GitHub обновляется ежедневно.
Похожие проекты
A self-hosted, OpenAI-compatible AI engine that runs LLMs, vision, voice, image and video models on any hardware, including CPU-only machines.
Python infrastructure that turns enterprise data into a graph-based context layer with ontologies, reasoning, provenance and decision audit trails for AI systems.
Incremental data pipeline framework in Python with a Rust core that keeps vector indexes and knowledge graphs in sync with changing sources for AI agents and RAG apps.
Parses multi-language codebases into a knowledge graph and lets you query, search, and edit the code in plain English via an AI agent.
A Python library and runnable Jupyter textbook implementing 35 agentic AI architectures, from Reflexion and ReAct to MemGPT and BrowserAgent, on LangGraph.