Salt la conținutul principal
Microsoft
separator
https://catalogartifact.azureedge.net/publicartifacts/bcloudllc1671615348068.graphrag122-9b6a49cf-e96f-40fd-bc69-89289191e22f/image3_bCloudlog.png

GraphRAG

de bCloud LLC

Version 3.1.2 + Free Support on Ubuntu 26.04

Microsoft GraphRAG 3.1.2 is a command-line artificial intelligence solution that converts unstructured text into structured knowledge graphs. It uses graph-based Retrieval-Augmented Generation (RAG) to identify entities, relationships, communities, and important information within documents.

The solution supports document indexing, knowledge-graph creation, community detection, semantic retrieval, and context-aware question answering. It is suitable for research, enterprise knowledge management, document analysis, and AI-powered information retrieval.

Features of Microsoft GraphRAG 3.1.2:

  • Creates structured knowledge graphs from text documents.
  • Identifies entities and relationships within uploaded information.
  • Supports local, global, DRIFT, and basic search methods.
  • Generates community summaries for large datasets.
  • Provides context-aware answers through supported AI models.
  • Supports OpenAI, Azure OpenAI, and compatible model providers.
  • Runs through a command-line interface on Ubuntu.

Usage instructions for Microsoft GraphRAG:
$ sudo su
$ cd /opt/graphrag
$ source .venv/bin/activate
$ cd /opt/graphrag/workspace
$ graphrag index
$ graphrag query --method local --query "Enter your question"

Configuration and API credentials are saved in: /opt/graphrag/workspace/.env

GraphRAG configuration is saved in: /opt/graphrag/workspace/settings.yaml

Access Microsoft GraphRAG:

Connect to the Azure virtual machine using SSH and run GraphRAG commands from the terminal.

Microsoft GraphRAG is a CLI application. It does not provide an official web dashboard and does not require a browser URL or application port.

Disclaimer: Microsoft GraphRAG 3.1.2 is provided “as is” under its applicable open-source license. Users are responsible for configuring supported AI providers, protecting API credentials, monitoring model usage costs, securing input data, and validating generated responses. AI-generated results may contain incomplete or inaccurate information and should be reviewed before production use.

Română (Republica Moldova)
Pictograma Renunțare la opțiunile de confidențialitate Opțiunile dumneavoastră de confidențialitate
Confidențialitatea pentru sănătatea consumatorilor Hartă de site Contactați-ne Confidențialitate și module cookie Condiții de utilizare Despre reclame Gestionați modulele cookie