title: "MCP +フェデレーテッドサーチはインデックスを強制終了しますか?"

The rise of federated search, driven by new connectors from open protocols like OpenAI and Google Agentspace, is prompting a re-evaluation of traditional data indexing. This raises the question of whether Large Language Models (LLMs) still require a pre-built index for information retrieval. The core debate centers on whether the real-time capabilities of federated search will render the practice of indexing obsolete for enterprise AI applications. This involves weighing the benefits of live data access against the performance and structure provided by a centralized index.