Gemini Knowledge Hub

A portable Python interface for Google's Gemini File Search API. Create queryable knowledge bases from local documents with citation support.

Features

• Store Management - Create/retrieve file search stores
• Document Ingestion - Recursive upload of PDF, TXT, MD, DOCX, and 30+ file types
• Semantic Search - Query documents with natural language
• Citations - Automatic source attribution in responses
• Hot-Reload - Add documents without recreating stores

Quick Start

1. Configure

Edit project_config.json:

{
  "gemini_api_key": "YOUR_API_KEY_HERE",
  "store_display_name": "My Knowledge Base",
  "documents": ["./documents"],
  "model": "gemini-2.5-flash"
}

2. Setup

python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install google-genai python-dotenv
python validate_setup.py
python setup_gemini.py

3. Query

python test_query.py

Project Structure

File	Purpose
setup_gemini.py	Core GeminiFileSearch class
validate_setup.py	Pre-flight environment check
verify_system.py	Verify existing stores
test_query.py	Query testing
project_config.json	User configuration

Requirements

• Python 3.8+
• google-genai (NOT google-generativeai)
• python-dotenv
• Google AI Studio API key

Documentation

• AI Setup Prompt - AI assistant instructions
• Guide - Full SDK reference & troubleshooting
• Use Cases - Example applications

License

N/A