The Model Context Protocol (MCP) has been updated to remove protocol-level sessions, requiring requests to be self-describing. This change, effective July 28, 2026, aims to improve scalability by allowing requests to be handled by any server instance without sticky routing. To manage state between calls, the new pattern involves explicit state handles, where tools mint and return visible handles that the model can pass around, offering benefits like multi-handle support, better debugging, and explicit scoping. A new CLI tool, stateforge, has been developed to generate TypeScript/Express servers implementing this pattern, with swappable state stores like Redis and deployment configurations for Vercel and Docker. AI
IMPACT Introduces a new pattern for managing state in AI model interactions, potentially improving scalability and developer experience.
RANK_REASON The item describes a new CLI tool and a pattern for managing state in a protocol, rather than a core AI model release or significant industry event.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
