Open Knowledge Tree
The source-driven knowledge engine.
Fetch a URL or DOI, and OKT extracts atomic facts, deduplicates them by meaning, links them into a concept graph, and accumulates syntheses you can search, cite, and reason over.
The Data Pipeline at a Glance
Deep Dive on System Mechanisms
The OKT paper itself — written by OKT. A research flow ran against OKT's own repository to synthesize the system's mechanisms into a single, fully-cited document. It is OKT showcasing its own research capabilities.
Atomic Facts and Emergent Concept Graphs
OKT self-showcaseThe OKT paper — written by OKT. A self-referential deep dive on OKT's own mechanisms: atomic fact decomposition, the emergent concept graph, and the fact-co-occurrence relations that fall out of it. A research flow ran against OKT's own repository to synthesize the system's mechanisms into a single document, with every claim traceable to a fact and its sources.
- Self-showcase: OKT researching itself, end-to-end, with its own agentic flow.
- Two citation kinds, visually differentiated — amber direct cites (researcher-chosen) vs blue auto-matched annotations (system-retrieved).
- 122 direct cites + 109 auto-matched annotations, every one clickable to its supporting fact and source URLs.
Live Examples
Real meta-syntheses produced by OKT's agentic flow. Click any citation to see the supporting fact and its sources. Each example header names the model that authored the synthesis.
The Modular Tropical Agroforestry Recipe Book
A 4-scope meta-synthesis (~1,300 sources, 100,000+ facts) integrating tropical polyculture architecture, belowground mechanisms, mycorrhizal networks, and mushroom & pest ecology. Every citation is clickable and carries a posture.
Human Alimentation: A Multidimensional Feeding Meta-Synthesis
A 9-scope meta-synthesis (499 facts) integrating global foodways, historical diets, protein sources, nutrient matrices, lifespan evidence, mood & cognition, life-stage physiology, lived evidence, and governance. Every citation is clickable and carries a posture.
Explore the Docs
Knowledge Flow
The 7-stage pipeline: source extraction, fact decomposition with coreference resolution, embedding-based dedup, concept graph, summaries, and synthesis.
MCP Tools
18 Model Context Protocol tools to fetch sources, search facts and concepts, track ingestion, and create annotated reports — all over OAuth 2.1.
REST API
HTTP endpoints for auth, repositories, sources, facts, concepts, investigations, reports, and providers. JWT and OAuth 2.1 bearer auth.
Develop from Source
Clone the repo and run `just dev` for hot-reload, source mounts, and the full dev toolchain with Go, Node, and e2e tests.
Architecture
Multi-database layout, Casbin RBAC, provider strategy, River task manager, Qdrant collections, and the full DB schema.