Release Notes#
Release 1.0.2#
Summary#
Latest release with bug fixes, Qwen3-VL support and documentation updates.
Changes#
Bug fixes
Qwen3-VL support
Documentation updates
Release 1.0.1#
Summary#
Minor release to add helm chart, documentation and config update and support llama 3.1 8B NIM
Changes#
Added helm chart
Documentation update
Config update
LLM support: LLaMa 3.1 8B
Release 1.0.0#
Summary#
Major release featuring enhanced retrieval capabilities, expanded database support, and improved modular configuration system.
New Features#
Enhanced Retrieval Methods#
Chain of Thought (CoT): Added reasoning-based retrieval for improved context understanding
Vision Language Model (VLM): Integrated multi-modal retrieval capabilities for processing visual content
Advanced Graph RAG (AdvGRAG): Implemented sophisticated graph traversal and retrieval algorithms for better performance
Expanded Database Support#
ArangoDB: Added support for graph-based document storage and retrieval
Elasticsearch: Integrated full-text search capabilities with distributed search support
Configuration Improvements#
Modular Configuration System: Redesigned configuration architecture for easier function and tool creation
Simplified Addition Process: Streamlined workflow for adding new functions and tools to the system
Enhanced Extensibility: Improved modularity enables faster development and deployment of new components
Experimental Features#
MCP Tools: Experimental MCP tools to enable AI agents interacting with video.
Structured Response: JSON Structured Response mode for responses from CA-RAG.
Release 0.5.0#
Summary#
This is the first general release of Context Aware RAG.