TrustGraph is an open-source AI provisioning platform designed to simplify the deployment and management of complex AI environments. It targets engineers and researchers building AI applications, offering a unified, environment-agnostic solution for deploying containerized AI tools, pipelines, and integrations, with a focus on secure and simplified operations.

How It Works

TrustGraph utilizes a modular architecture centered around Apache Pulsar as a pub/sub backbone for data processing queues and module communication. Its core innovation is TrustRAG, an advanced Retrieval Augmented Generation (RAG) approach that constructs knowledge graphs from raw data. This involves extracting entities and relationships, mapping them to vector embeddings, and then performing hybrid retrieval. Queries trigger subgraph traversal within the knowledge graph to generate rich, context-aware prompts for LLMs, enhancing accuracy and relevance beyond traditional text chunking.

Quick Start & Requirements

• Install the CLI: pip3 install trustgraph-cli==0.21.17 (CLI version must match release version).
• Deployment: Use the Configuration Builder to generate YAML files, then deploy with docker compose up -d.
• Prerequisites: Docker is recommended for initial setup.
• Resources: Deployment configurations are available for Docker, Podman, Minikube, AWS, Azure, Google Cloud, and Scaleway.
• Links: Full Docs, Configuration Builder, API Docs, CLI Docs.

Highlighted Details

• Supports a wide range of LLM providers (OpenAI, Anthropic, Google, Mistral, etc.) and vector databases (Qdrant, Pinecone, Milvus).
• Features TrustRAG for enhanced RAG with knowledge graph traversal and reusable "Knowledge Cores."
• Includes a Test Suite for graph RAG chat, vector search, and data loading.
• Offers automated data extraction agents for building knowledge graphs from PDFs, text, and Markdown.

Maintenance & Community

• Active community support via Discord.
• Regular updates and releases indicated by PyPI versioning.
• Links: Discord, YouTube.

Licensing & Compatibility

• Licensed under AGPL-3.0.
• AGPL-3.0 is a strong copyleft license, requiring derivative works to also be open-sourced under the same license. This may impose restrictions on linking with closed-source commercial applications.

Limitations & Caveats

The AGPL-3.0 license may restrict commercial use or integration into proprietary systems without open-sourcing the entire application. The project is actively developed, and specific component versions might require careful management to ensure compatibility.