Google Rolls Out Interactions API for Next-Gen AI Agents

December 11, 2025

Details

  • Google has launched the Interactions API in public beta as of December 11, 2025, offering developers a single interface to build agentic applications leveraging Gemini models and AI agents.
  • The API, accessed via Google AI Studio with existing Gemini API keys, features one RESTful endpoint for direct model use and agent workflows, starting with Gemini Deep Research as the inaugural built-in agent.
  • Main features include server-side state management for conversation histories, background execution for handling lengthy tasks, and native Model Context Protocol (MCP) server support allowing models to use external tools and data sources seamlessly.
  • Gemini Deep Research showcases the platform’s capabilities as a research agent that can perform deep investigations and generate detailed reports—marking a shift from Google's prior stateless generateContent API toward more dynamic, stateful interactions.
  • The Interactions API is part of a larger push involving the Agent Development Kit (ADK) and Agent2Agent (A2A) protocol, with plans for integration into Vertex AI to support enterprise-scale adoption.

Impact

With this launch, Google signals a decisive move toward agent-centric AI development, making it easier for developers to build robust, production-ready agents. The integrated server-side state and tool orchestration align with the industry's broader evolution toward intelligent, autonomous systems. Upcoming Vertex AI support is poised to strengthen Google's competitiveness among enterprise customers seeking scalable AI solutions.

Read more