What is TrustGraph Engine
TrustGraph Engine is a knowledge agent development platform that does not depend on a specific large language model (LLM). It provides tools, services, graph databases, and vector databases to deploy reliable, scalable, and accurate AI agents. TrustGraph Engine features batch document ingestion, automated knowledge graph construction, automated vectorization, model-agnostic LLM integration, RAG combining knowledge graphs and vector databases, enterprise-grade reliability, scalability, and modularity, as well as data privacy protection.
Main features of TrustGraph Engine
- Batch document ingestion: Can handle the ingestion of large numbers of documents.
- Automated knowledge graph construction: Automatically constructs knowledge graphs, simplifying knowledge management.
- Automated vectorization: Automatically converts text into vector form for easier processing.
- Model-agnostic LLM integration: Supports integration of different models, not limited to a specific LLM.
- RAG combining knowledge graphs and vector databases: Enhances retrieval capabilities based on knowledge graphs and vector databases.
- Enterprise-grade reliability, scalability, and modularity: Meets enterprise-level application requirements.
- Data privacy protection: Supports local LLM deployment, such as Ollama and Llamafile, enhancing data privacy.
Technical principles of TrustGraph Engine
- Modular architecture: Designed to be modular, supporting multiple language models and environments.
- Publish/subscribe backbone: Based on Apache Pulsar as the pub/sub backbone, connecting different processing modules.
- Automatic knowledge agents: Uses three automatic knowledge agents to extract knowledge from text corpora (PDF or text) into a hyper-dense knowledge graph, including a topic extraction agent, an entity extraction agent, and a node connection agent.
- Containerized deployment: TrustGraph is fully containerized, deployed locally based on Docker, Podman, or Minikube, and can be deployed on Google Cloud using Kubernetes.
- YAML configuration files: All necessary configurations are built into YAML files, and each model deployment and graph storage configuration has a corresponding YAML configuration file.
- Data flow management: Pulsar serves as the backbone of the data flow, managing data queues connected to processing modules.
Project address of TrustGraph Engine
- Project official website：trustgraph.ai
- GitHub repository：https://github.com/trustgraph-ai/trustgraph
Application scenarios of TrustGraph Engine
- Enterprise data analysts: Need to extract information from large numbers of documents and build knowledge graphs to support decision-making.
- Researchers and scholars: In the academic field, need to process and analyze large amounts of research materials and build domain knowledge graphs.
- Legal professionals: In the legal field, need to conduct in-depth analysis of legal documents and discover connections between cases.
- Financial analysts: In the financial industry, need to analyze financial reports and market data and build risk assessment models.
- IT and software developers: Technical personnel who need to integrate and develop knowledge graph-based applications, or need to customize and extend AI solutions.
- Security analysts: In the field of cybersecurity, it is necessary to analyze log files and network traffic to identify potential security threats.