n8n-nodes-gemini-search

n8n-nodes-gemini-search

This is an n8n community node package that provides nodes to interact with Google's Gemini AI models for search and content generation capabilities.

Features

• Gemini Search Node: A comprehensive node that can perform web searches and generate content using Google's Gemini models
• Gemini Search Tool: A simplified tool version optimized for quick searches with Google Gemini
• URL Restriction: Limit searches to specific websites or domains
• Organization Context: Provide company or domain-specific context for more relevant results
• Multiple Model Support: Choose from various Gemini models including latest versions (1.0 Pro, 1.5 Pro, 2.0 Flash, 2.0 Pro, 2.5 Pro)

Installation

In n8n Desktop

• Go to Settings > Community Nodes
• Select "Install"
• Enter n8n-nodes-gemini-search in the "npm package name" field
• Click "Install"

In n8n Cloud

Use the Community Nodes feature to add this node to your cloud instance.

Usage

Gemini Search Node

This node offers two operations:

• Web Search: Performs a web search using Gemini's capabilities
• Generate Content: Generates content using Gemini without web search

Configuration Options:

• API Key: Google Gemini API key (required)
• Model: Select from available Gemini models
• Prompt/Query: Your search query or content generation prompt
• Organization Context: Optional context to focus the search on a specific organization
• URL Restrictions: Limit search to specific websites (comma-separated)
• System Instruction: Custom instructions to guide the AI behavior
• Advanced Options:
  • Temperature: Control response randomness (0.0-1.0)
  • Max Output Tokens: Limit response length
  • Top P/Top K: Fine-tune token selection
  • Extract Source URL: Retrieve source URLs for search results

Gemini Search Tool

This is a simplified version focused on search, optimized for use as an n8n tool:

• Query: Enter your search query
• Model: Select from available Gemini models
• Organization Context: Optional context to focus on specific organizations
• URL Restrictions: Limit search to specific websites
• Options:
  • Temperature: Control response randomness
  • Max Output Tokens: Limit response length
  • Custom System Instruction: Override default instructions
  • Return Full Response: Include complete API response data
  • Extract Source URL: Retrieve source URLs

Example Workflows

Basic Search

• Create a new workflow
• Add a "Gemini Search Tool" node
• Configure your API credentials
• Enter a search query
• Execute and view the search results

Advanced Content Generation

• Create a new workflow
• Add a "Gemini Search" node
• Set operation to "Generate Content"
• Configure model and prompt
• Adjust temperature and token settings as needed
• Execute to generate AI content

Using as a Tool

This node can be used as a tool in n8n AI Agents. To enable community nodes as tools, you need to set the N8N_COMMUNITY_PACKAGES_ALLOW_TOOL_USAGE environment variable to true.

Setting the Environment Variable

Bash/Zsh Shell

export N8N_COMMUNITY_PACKAGES_ALLOW_TOOL_USAGE=true
n8n start

Docker (docker-compose.yml)

environment:
  - N8N_COMMUNITY_PACKAGES_ALLOW_TOOL_USAGE=true

Desktop App

Create a .env file in the n8n directory:

N8N_COMMUNITY_PACKAGES_ALLOW_TOOL_USAGE=true

Permanent Setting (Mac/Linux)

Add to your ~/.zshrc or ~/.bash_profile:

export N8N_COMMUNITY_PACKAGES_ALLOW_TOOL_USAGE=true

Credentials

This node requires a Google Gemini API key. You can get one from the Google AI Studio.

Support

For issues or feature requests, please open an issue on GitHub.

License

MIT

v1.1.10

This version introduces dynamic model loading and the new Gemini URL Context tool.

Key Updates:

• Dynamic Model Loading: Models are now fetched directly from the Gemini API, ensuring the list is always up-to-date.
• URL Context Tool: Leverages Gemini's new capability to analyze content from URLs provided in the prompt. This is available in the "Generate Content" operation of the Gemini Search node and is enabled by default in the Gemini Search Tool node.
• Fixes & Improvements: Includes fixes for linting issues and general code refinements.