Skip to main content

​
Overview

Gemini is Google’s AI coding agent. When configured in Zenflow, it runs inside isolated Git worktrees with access to Zencoder’s context engine, multi-repo search, and Zen Rules. Gemini supports multi-modal inputs including images and screenshots.

​
Setup

​
Prerequisites

You need one of:
  • Google AI API key: Get one at ai.google.dev
  • Vertex AI access: For enterprise deployments

​
Installation

1

Install Gemini CLI

Install the Gemini agent on your machine. See Google’s documentation for installation instructions.
2

Authenticate

Configure your Google AI API key or Vertex AI credentials.
3

Select in Zenflow

In Zenflow, go to Settings → Saved presets and select Gemini as the agent for your tasks.

​
When to Use Gemini

  • Multi-modal tasks: Implement UIs from screenshots or mockups
  • Large context windows: Work with very large codebases
  • Design-to-code: Convert Figma exports into working components
  • Visual debugging: Debug rendering issues from screenshots

​
Configurations

  • Model – default for higher quality, flash for faster responses.

​
Pricing

OptionHow It Works
Google AI APIPay-per-use with a free tier
Vertex AIEnterprise pricing with SLAs

​
Next Steps

Supported Agents

Compare all available agents

Figma to Code

Design-to-code tutorial