A specialized tool that leverages Google AI Pro subscriptions to generate and edit images through the Gemini Web API without requiring standard API keys.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install gemini-web-image
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install gemini-web-image using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Gemini Web Image?
Gemini Web Image is a robust utility designed to integrate with Openclaw Skills, enabling developers to harness the power of the Gemini 3 Pro Image model. By utilizing browser-based authentication rather than traditional API billing, it provides a cost-effective and powerful way to generate illustrations, diagrams, and concept art directly from your local environment. This skill is particularly valuable for those who already maintain a Google AI Pro subscription and want to automate their visual creative workflows.
The tool is built with a focus on ease of use and security. It utilizes local browser cookies to authenticate with gemini.google.com, ensuring that your credentials remain on your machine while providing full access to Gemini's generative capabilities. Whether you are building product mockups or visualizing complex ideas, this skill bridges the gap between web-based AI interfaces and local development scripts.
Gemini Web Image Use Cases
- Creating high-fidelity illustrations and concept art from natural language descriptions.
- Generating technical diagrams and product mockups for presentations or documentation.
- Automating image generation tasks within larger AI agent workflows using Openclaw Skills.
- Performing instruction-based edits on existing images to iterate on visual designs.
How Gemini Web Image Works
- The skill initiates by accessing your Google AI Pro session via Chrome's local cookie store or a manual JSON configuration.
- When a prompt is received, the script uses uv to manage dependencies and execute the generation command.
- The request is transmitted to the Gemini Web API, simulating a browser-based interaction to generate the image.
- Upon receipt, the skill can automatically apply watermark removal using a lossless Reverse Alpha Blending technique.
- The final image is saved locally with a timestamped filename, and a media reference is passed back to the Openclaw Skills ecosystem for immediate use.
Gemini Web Image Setup
First, ensure you have a Google AI Pro subscription and are logged into Chrome. Install the uv package manager:
brew install uv
For most users, the skill will automatically detect your session. For headless or server environments, you can manually provide cookies:
mkdir -p ~/.config/gemini
# Populate with secure_1psid and secure_1psidts from Chrome DevTools
cat > ~/.config/gemini/cookies.json << 'EOF'
{"secure_1psid": "YOUR_VALUE", "secure_1psidts": "YOUR_VALUE"}
EOF
chmod 600 ~/.config/gemini/cookies.json
Gemini Web Image Data Schema & Taxonomy
| Data Entity | Details |
|---|---|
| Authentication | Uses ~/.config/gemini/cookies.json for manual session tokens. |
| Image Output | Saves files as .png using a yyyy-mm-dd-hh-mm-ss-name.png format. |
| Dependencies | Managed by uv, including gemini-webapi, Pillow, and numpy. |
| Watermark Refs | Utilizes bg_48.png and bg_96.png as reference layers for cleaning. |
Gemini Web Image Advanced Features
- Lossless watermark removal using specialized Reverse Alpha Blending reference images.
- Support for image-to-image editing by providing an input file alongside instructions.
- Conversation management with the
--keep-chatflag to persist threads on the Gemini web interface. - Headless mode support for integration into remote server environments and automated pipelines.
- Automated dependency resolution via uv, requiring no manual pip installations for Openclaw Skills users.
SKILL.md
Loading
Related Openclaw Skills
Felo AI Search
A powerful AI-synthesized search skill that aggregates and summarizes information from up to 40 web sources for deep research.
Geospatial OSINT Dashboard
A comprehensive framework for building real-time geospatial intelligence dashboards with 3D rendering and tactical visual effects.
Ad Symbiont: Build Amazon Affiliate WordPress Plugin
An automated workflow for generating a professional WordPress plugin that transforms Amazon links into high-converting affiliate ad units.
Clawgora
Clawgora is a dedicated labor marketplace that enables AI agents to post jobs, outsource complex tasks, and earn credits by fulfilling work for other agents.
OpenClaw Skill Creator
A specialized tool designed to help developers create, evaluate, and refine high-performance Openclaw Skills for AI agents.
YouTube Transcribe
A high-performance transcription skill that intelligently switches between YouTube captions and local AI Whisper models for maximum accuracy.
Featured*
