VeritasGraph
VeritasGraph — open-source Knowledge Graph & GraphRAG framework on GitHub. Build multi-hop reasoning, ontology-aware retrieval, and verifiable attribution over
What is VeritasGraph?
VeritasGraph — open-source Knowledge Graph & GraphRAG framework on GitHub. Build multi-hop reasoning, ontology-aware retrieval, and verifiable attribution over your own data. Nodes, edges, RDF, linked-data — runs locally or in the cloud.
Overview of VeritasGraph
What it is
VeritasGraph is an open-source Knowledge Graph and GraphRAG framework. It combines tree-search structure with knowledge-graph reasoning to build governed AI agents. The framework runs 100% locally or in the cloud and supports ontology-aware retrieval and verifiable attribution.
Key features
- Combines tree-search structure with knowledge-graph reasoning
- Supports multi-hop reasoning and ontology-aware retrieval
- Verifiable attribution: stores provenance at write time to prevent hallucinated citations
- Deterministic retrieval via graph traversal, separating factual evidence from LLM generation
- Explicit abstention: agent identifies missing graph edges and returns “insufficient evidence”
- Runs 100% locally or in the cloud; supports Ollama, Neo4j, and cloud APIs
- Includes Studio local Agent Build Workspace with knowledge graph builder, agent workspace, guardrails, and KPI dashboard
Use cases
- Multi-hop reasoning over structured data
- Fact-checking and evidence-focused explanation
- Enterprise GraphRAG and verifiable AI platform
- Medical and pharmaceutical workflows (e.g., MedTrust Agent)
- Construction drawing takeoff (e.g., Floorplan-KG demo)
Getting started
bashpip install veritasgraphveritasgraph demo --mode=liteOr see the repository README for full instructions.
License and pricing
License: MIT. The framework is open-source and priced as open-source.
Based on the project's public README and repository metadata.