SAN FRANCISCO, CA – August 22, 2025 – TrustGraph, the open-source AI Product Creation Platform, today announced the major release of TrustGraph 1.2. This landmark update empowers AI agents to autonomously build and populate knowledge graphs, introduces foundational support for structured data, and expands its ecosystem with new integrations, including Anthropic models on Google VertexAI and deployment on OVHcloud.
The centerpiece of version 1.2 is the new Knowledge Extraction via Agent Framework. Leveraging the powerful ReAct framework, AI agents within TrustGraph can now do more than just query existing data; they can actively reason, use tools, and extract information to construct and enrich the knowledge graph themselves. This shifts TrustGraph from a system that processes data to one that enables AI to actively build its own deep contextual understanding.
“With version 1.2, we're taking a major step towards truly autonomous AI systems,” said Mark Adams, Co-Founder of TrustGraph. “Our new agent framework doesn't just answer questions about data; it actively builds the contextual understanding required for deep reasoning. By adding structured data support and expanding our LLM and cloud options, we are unlocking the full spectrum of enterprise knowledge for these next-generation intelligent agents.”
Key highlights of the TrustGraph 1.2 release include:
AI Agent-Powered Knowledge Extraction: A new kg-extract-agent utilizes the ReAct framework, enabling agents to autonomously populate the knowledge graph, representing a significant leap in AI capabilities. Agent tool support has also been enhanced for more complex, multi-step reasoning.
Structured Data Handling (MVP): For the first time, TrustGraph can ingest and process structured data. This initial implementation includes new schemas, an object extractor, and Cassandra storage support, allowing agents to reason over a much broader range of enterprise data sources beyond unstructured documents.
Expanded LLM and Cloud Support: The platform now supports Anthropic models on Google VertexAI, giving users more choice in powerful large language models. Additionally, new deployment guides for OVHcloud expand the options for hosting TrustGraph in production and development environments.
New Developer and Operator Tooling: A new command-line utility for loading knowledge (Knowledge Load Utility CLI) simplifies data ingestion, making it easier than ever to get started and manage knowledge within the system.
Enhanced Stability and Modernization: The release includes a comprehensive new logging strategy for improved observability, a migration to the modern pyproject.toml build system, and numerous bug fixes that improve the resilience of agent tools, data validation, and core components.
This release solidifies TrustGraph's position as a leading platform for building sophisticated AI products with deep contextual awareness, providing the transparency and power of knowledge graphs with the advanced reasoning capabilities of modern agent frameworks.
TrustGraph 1.2 is available immediately for download and deployment. For more information, visit the TrustGraph documentation.
About TrustGraph
TrustGraph is an Open Source AI Product Creation Platform that helps organizations build, deploy, and manage sophisticated AI agents with deep contextual understanding. Unlike traditional AI systems that work with isolated data points, TrustGraph creates interconnected Knowledge Graphs from enterprise data, enabling agents to understand relationships and context. By combining its advanced context engineering technology with powerful AI agents, TrustGraph solves common AI challenges like hallucinations and lack of context, providing radical transparency and unifying siloed organizational data into coherent, actionable knowledge systems.
Contact:
Discord Community
https://trustgraph.ai
https://github.com/trustgraph-ai/trustgraph