Does this require setting up a separate vector database?

No, Gemini File Search is a fully managed RAG service that handles chunking, embeddings, and vector search internally without needing external databases like Pinecone or Vectorize.

Which models are supported by the Gemini File Search skill?

This skill specifically requires Google Gemini 2.5 Pro or Gemini 2.5 Flash models; it is not compatible with older versions such as Gemini 1.5.

How does storage pricing work for this skill?

While indexing costs roughly $0.15 per 1M tokens, storage is typically free within the 10GB-1TB limits depending on your tier, though embeddings and metadata can triple the original file size.

What file formats are supported?

It supports over 100 formats including PDFs, Word documents (.docx), Excel spreadsheets (.xlsx), Markdown, JSON, and most common programming languages.

Can I update an existing document in the store?

No, documents in the Gemini File Search store are immutable. To update a file, you must follow the delete and re-upload pattern outlined in the skill documentation.

Google Gemini File Search

Name: Google Gemini File Search
Author: secondsky

bysecondsky

•

数据科学与机器学习

Integrates Google Gemini's managed RAG capabilities to provide high-performance document Q&A and knowledge base search across 100+ file formats.

This skill enables Claude Code to leverage Google Gemini's fully managed Retrieval-Augmented Generation (RAG) system, eliminating the requirement for manual vector database setup. It supports a vast array of document types—including PDFs, codebases, and spreadsheets—providing built-in grounding and citation metadata for highly accurate technical support and knowledge retrieval. Designed for production environments, it features automatic chunking, cost-effective indexing, and specialized support for Gemini 2.5 Pro and Flash models, making it ideal for developers who need a standalone, reliable document search solution.

主要功能

01Fully managed RAG with automatic chunking and embedding generation

02Cost-effective indexing at $0.15 per million tokens with free storage tiers

03Support for 100+ file formats including PDF, Excel, and various code files

04Built-in citation metadata for grounded and verifiable AI responses

0521 GitHub stars

06Seamless integration with Gemini 2.5 Pro and Flash model architectures

使用场景

01Executing complex Q&A across large internal documentation sets or legal files

02Indexing entire code repositories for context-aware developer assistance

03Building a searchable customer support knowledge base from technical manuals

What are Skills?·How to Install

Install with 🐟 Skill.Fish

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

For use in Claude.ai and ChatGPT

Download Skill