How many file formats can I upload for indexing?

The API supports over 100 different file formats, including PDF, Microsoft Office documents (Word, Excel, PowerPoint), and a wide variety of source code files.

What is the recommended chunking configuration?

For technical documentation, we recommend 500 tokens per chunk with a 50-token overlap to ensure optimal retrieval precision and context retention.

What are the costs associated with Gemini File Search?

Indexing costs $0.15 per 1 million input tokens as a one-time fee. Storage is free within specific tier limits, such as 10 GB for Tier 1 users.

Which models support Gemini File Search?

File Search is currently supported exclusively by Gemini 2.5 Pro and Gemini 2.5 Flash models. It is not compatible with Gemini 1.5 versions.

How do I update a document that has already been indexed?

Documents in Gemini File Search are immutable. To update a document, you must use the delete-and-re-upload pattern, which this skill automates for you.

Google Gemini File Search

Name: Google Gemini File Search
Author: jezweb

byjezweb

•

Ciencia de Datos y ML

Integrates Google Gemini's managed RAG system to build searchable document knowledge bases with automatic chunking, embeddings, and citations.

This skill enables Claude Code to implement and manage Google Gemini File Search, a powerful Retrieval-Augmented Generation (RAG) solution. It streamlines the process of uploading and indexing over 100 file formats—including PDFs, Word documents, and source code—into semantic search stores while handling complex backend requirements like automatic embeddings and citation generation. Designed for developers building document Q&A systems, the skill provides built-in protections against common API pitfalls such as document immutability errors, storage quota mismanagement, and improper chunking configurations, ensuring high-quality retrieval using Gemini 2.5 Pro and Flash models.

Características Principales

01Support for 100+ file formats including PDF, Excel, Word, and source code

02Optimized chunking strategies for technical, legal, and prose content types

03Fully managed RAG with automatic chunking and vector embedding generation

04Automated error prevention for storage quotas, metadata limits, and immutability

0591 GitHub stars

06Pre-configured integration patterns for Cloudflare Workers and Next.js

Casos de Uso

01Implementing semantic search across large-scale technical documentation repositories

02Developing AI applications that require verifiable citations from source documents

03Building internal company knowledge bases and AI-powered document Q&A systems

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jezweb/claude-skills google-gemini-file-search

For use in Claude.ai and ChatGPT

Download Skill