Google has open-sourced a full-stack research agent stack powered by its Gemini 2.5 AI model and LangGraph. This open-source quickstart enables developers to build intelligent research assistants capable of dynamic multi-step web searches, reflection on results, and synthesis of comprehensive answers with citations. The stack features a user-friendly interface with streaming capabilities and is built using React.js and LangChainAI. While it is not officially confirmed to match Gemini’s production backend, the implementation is described as strikingly close. Additionally, a two-hour workshop has been made available to guide beginners through using Gemini 2.5, covering API access, multimodality, function calling, and more. The open-source release aims to facilitate deep research applications and has garnered attention for its potential in future developments.