RubyLLM 2.1 (rubyllm.com)

RubyLLM 2.1 has introduced a suite of significant enhancements, most notably the inclusion of a Model Context Protocol (MCP) client, which allows seamless connection to various server APIs like GitHub and Slack. This update enables developers to create agents that can not only evaluate and interact with different tools and APIs in real-time but also manage conversations across multiple providers dynamically. For instance, conversations can now switch between different models mid-discussion, enhancing flexibility and utility. The addition of typed judgments with TypeSafe's Jev models allows for structured data evaluations, further integrating decision-making capabilities into applications. The performance improvements in RubyLLM 2.1 are striking. The update optimizes processing speed by streamlining request handling, allowing concurrent calls to operate more efficiently, often reducing response times significantly compared to version 2.0. For example, streaming delays for data-rich tasks such as event parsing and API interactions have decreased dramatically. This version also introduces features like real-time progress reporting for lengthy operations, improved cache handling with customizable time-to-live settings, and enhanced attachment management, all contributing to a more responsive and resource-efficient experience for developers in the AI/ML community.