trustgraph
The semantic intelligence layer for ontologies — governed, traceable knowledge for AI agents
Utopia is an open-source rag & knowledge project by deeplethe (deeplethe/utopia on GitHub). It has 8,372 GitHub stars and 1,159 forks, is released under the Apache-2.0 license, and was last updated on GitHub on 15 October 2026.
Utopia — an open-source enterprise world model: a bitemporal knowledge substrate that learns from your documents and gives agents a governed, auditable memory
Clone the repo and bring the stack up with Docker Compose — the first account you register becomes the administrator.
Under Administration → Models, set the chat and embedding endpoints Utopia should use.
Start extracting business documents into the public knowledge base and watch the ontology evolve with time awareness.
Knowledge carries both valid time and record time, so past states of understanding can be replayed and audited.
Conflict detection, reasoning and decision support run against a governed ontology, not raw chunks.
Prebuilt images for x86-64 and ARM64 (Apple Silicon, Graviton) run entirely on hardware you control.
The roadmap adds checking agent calls against ontology rules and symbolic logic, plus agent memory over MCP.
Requirements: Docker; local development also needs Rust 1.85+, Node 20+ and pnpm.
git clone https://github.com/deeplethe/utopia.git
cd utopiaBrings up the app profile; then open http://localhost:1516 and register.
docker compose --profile app up -dUses the build overlay of the same Compose stack.
docker compose -f docker-compose.yml -f docker-compose.build.yml --profile app up -d --buildDeepLethe; the project site is utopia.bi.
Yes, Apache-2.0 licensed, with prebuilt container images on GHCR.
You configure your own chat and embedding model endpoints; Utopia itself has no licence fee.
A vector store retrieves similar chunks of present knowledge. Utopia models knowledge bitemporally over an ontology, so it can show what was believed, when it changed, and what a decision rested on.
Clone the repo, run docker compose --profile app up -d, open http://localhost:1516, register the admin account, and configure your model endpoints.
The semantic intelligence layer for ontologies — governed, traceable knowledge for AI agents
A private AI platform for agents, assistants and enterprise search over your documents
The open-source, permission-aware context layer that lets agents search all your company knowledge.
Skills that let coding agents search and read your local documents.
A knowledge-base platform with visual AI workflows — build Q&A systems on your own documents
Knowhere turns documents into structured, hierarchy-native chunks ready for agents and RAG.
The most-starred other RAG & Knowledge projects in the LibreAI catalogue — the realistic shortlist if Utopia is not the right fit.
MarkItDown — Microsoft tool that converts PDFs, Office files, images and more to clean Markdown for LLMs, MIT
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
Nearby RAG & Knowledge projects in the catalogue, closest in stars to Utopia — keep exploring the shelf either side of it.
Deep Lake — a multimodal datalake and vector store for AI, with version control and training streaming.
DeepSearcher — open-source deep research over your private data, with reasoning models and Milvus retrieval.
PaperQA2 — high-accuracy RAG for answering questions from scientific papers, with citations you can check.
Unstract — LLM-driven extraction of unstructured data, built for API deployments and ETL pipeline workflows.
Polyglot document intelligence in Rust — extract text, tables and structured data from 106 formats.
Vespa — the AI search platform: vector, text and tensor search at production scale.
Free and open — grab the source, star the repo, and start creating.