Today, we introduce the File Search Tool, a fully managed Retrieval-Augmented Generation (RAG) system integrated into the Gemini API. This tool simplifies your development process by abstracting the retrieval pipeline, allowing you to concentrate on building high-quality applications. With File Search, you can easily ground Gemini with your data, yielding accurate, relevant, and verifiable responses.
For cost-effective use, we offer free storage and query-time embedding generation, charging only $0.15 per million tokens when indexing files. Designed for seamless integration, the File Search Tool manages file storage, chunking, embeddings, and context injection within the generateContent API. It utilizes the advanced Gemini Embedding model for powerful vector searches, ensuring comprehensive results even with varied language. Additionally, built-in citations enhance response verification. The tool supports multiple formats, including PDF, DOCX, and JSON. Experience its capabilities through our demo app in Google AI Studio (requires a paid API key).
