Skip to content
BeeAI Framework

BeeAI Framework Community Update

August 26, 2025

Section titled “August 26, 2025”

🛠️ Improvements

Section titled “🛠️ Improvements”
  • Streaming - Better text and tool calls streaming processing and error handling
  • Parallel Tool Calls - Execute multiple tools simultaneously for better performance
  • Serve Module - MCP, BeeAI platform and A2A are now compatible with the latest versions
  • OpenTelemetry - Is more reliable and supports all modules in the framework (see our Observability docs page).

RAG Module

Section titled “RAG Module”
  • Vector Store Tool - Semantic search capabilities for any agent
  • Document Loader - Support for multiple document formats
  • Text Splitter Backend - Intelligent document chunking strategies
  • Dynamic Loading - Load RAG components using configuration strings

See the RAG docs page for more details.

🔧 Quality & Reliability Updates

Section titled “🔧 Quality & Reliability Updates”

LLMs, Agents and Tools

Section titled “LLMs, Agents and Tools”
  • Enhanced MCP client session cleanup
  • Add Google Gemini Backend Adapter
  • Fixed streaming chunk processing for better real-time responses
  • Enhanced tool choice validation and error messages
  • Add cost tracking (part of the ChatModelOutput)
  • Handle duplicated rules in RequirementAgent
  • Improve the performance of the Handoff tool

🗺️ What’s Coming Next

Section titled “🗺️ What’s Coming Next”

Priorities:

  1. Runnables Migration - Major refactoring to the agents so that are easier to use, compose and build.
  2. Workflows V2 - Introducing a decorator-based, fully declarative approach to workflows
  3. Loader Module - Configure frameworks’ primitives with YAML (JSON) files
  4. Agent Consolidation - Consolidate the agents into a single agent
  5. Built-in serving functionality

💬 Community Input Needed

Section titled “💬 Community Input Needed”
  • Workflows V2 - Let us know your thoughts on the proposal!
  • Chat Completion Serve Integration Let us know if you’d like to see the integration in the serve module.

Join the discussion: discord.gg/NradeA6ZNF

🙏 Community Recognition

Section titled “🙏 Community Recognition”

Core Maintainers

Section titled “Core Maintainers”

Thank you to our dedicated maintainers for their contributions this period:

  • @araujof - Implemented the foundational Runnable interface (#982)
  • @Tomas2D - Enhanced serialization (#1001), parallel tool calls (#986), streaming improvements (#973), schema transformations (#996), and multiple bug fixes
  • @xjacka - Built memory manager for serving (#983), A2A upgrades (#951), context-based memory (#975), and Watsonx integration improvements (#948)
  • @antonpibm - Developed RAG module: vector store tool (#991), document loader (#962), text splitter (#974), dynamic processors (#979), and comprehensive documentation (#952)

External Contributors

Section titled “External Contributors”

Special thanks to our community contributors this period:

  • @richardesp - Cost tracking and LiteLLM response cost integration (#926)
  • @nforro - Google Gemini backend implementation (#939)
  • @Niamorine - Fixed critical MCP tool error (#922)

Linux Foundation AI & Data

Section titled “Linux Foundation AI & Data”

BeeAI Framework is proud to be part of the Linux Foundation AI & Data program, fostering open, collaborative, and community-driven development.

🙏 Thank You

Section titled “🙏 Thank You”

The BeeAI Framework is community-driven - your feedback shapes our roadmap!

Questions? Let’s discuss!