A comprehensive developer assistant for integrating Google Gemini multimodal models with ultra-long context windows and native media understanding.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install gemini-guide
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-guide 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 Guide?
Gemini Guide is an expert-level technical resource designed to streamline the integration of Google's Gemini family of models into modern development workflows. As a core part of the Openclaw Skills library, this guide provides the necessary logic and patterns to utilize Gemini Pro and Flash models, which offer industry-leading context windows of up to 2 million tokens. It emphasizes the use of native multimodal capabilities, allowing developers to process text, images, video, and audio through a single unified API.
By incorporating this skill into your environment, you can build sophisticated AI agents that outperform standard LLMs in data-heavy tasks. The Openclaw Skills ecosystem ensures that you can leverage these powerful models while maintaining high performance and cost efficiency through advanced features like context caching and optimized model selection.
Gemini Guide Use Cases
- Analyzing massive codebases or entire libraries using the 2-million token context window.
- Building native video analysis tools that can summarize and query video content without frame extraction.
- Developing low-latency, cost-effective chatbots using Gemini Flash for high-concurrency environments.
- Implementing complex tool-use scenarios with robust function calling and JSON output modes.
- Optimizing operational expenses for recurring large-document queries through persistent context caching.
How Gemini Guide Works
- Configure your environment by obtaining an API key from Google AI Studio and installing the google-genai SDK.
- Select the appropriate model variant based on whether you need maximum reasoning (Pro), speed (Flash), or cost-efficiency (Flash-lite).
- Prepare input content which can consist of mixed media types, including raw file bytes or cloud-stored URIs for video and audio.
- Define optional tools and functions if your implementation requires the agent to interact with external systems or APIs.
- Send the request to the Gemini API and handle the response, utilizing Openclaw Skills patterns for structured data parsing and error handling.
Gemini Guide Setup
To integrate this skill with your Openclaw Skills setup, follow these installation steps:
pip install google-genai
Ensure your API key is exported as an environment variable or passed directly to the GenAI client during initialization. Access your keys at https://aistudio.google.com/apikey.
Gemini Guide Data Schema & Taxonomy
The skill manages data across several key structures to ensure reliable model interactions:
| Component | Type | Description |
|---|---|---|
Part |
Object | The fundamental unit of input (text, image, video, or audio) |
Context Cache |
Resource | Persistent storage of tokenized content for repeated access |
FunctionDeclaration |
Schema | Metadata defining the parameters and purpose of external tools |
GenerateContentConfig |
Config | Settings for temperature, response types (JSON/Text), and safety filters |
File |
URI/Bytes | Managed media assets for multimodal processing |
Gemini Guide Advanced Features
- Direct video and audio file uploads for native multimodal inference.
- Context Caching which significantly reduces costs (up to 75%+) for large-scale repeated queries.
- Native JSON mode to ensure that Openclaw Skills integrations receive perfectly formatted data for downstream tasks.
- Support for Gemini 2.0 Flash-lite, offering the highest throughput for simple, high-frequency automation.
- Flexible tool calling mechanisms that allow models to suggest and execute external logic.
SKILL.md
Loading
Related Openclaw Skills
Academic Research Hub
A centralized power-tool for academic discovery that searches multiple databases, manages citations, and aggregates research materials for systematic reviews.
Supply Chain Optimization for TikTok Shop
A specialized diagnostic tool designed for TikTok Shop sellers to identify supply chain bottlenecks, optimize affiliate payouts, and manage viral product lifecycles.
Content Workflow
A comprehensive content production pipeline that coordinates specialized AI agents to move from topic ideation to final Markdown publication.
Go Web Expert
A comprehensive Go development persona that enforces zero global state, explicit error handling, and strict testability for production-grade web applications.
Jina Reader & Search Skill
A powerful utility to fetch web pages as LLM-friendly markdown and perform structured web searches using Jina AI APIs.
TikTok Profile Scraper
A comprehensive browser-based tool for discovering and scraping TikTok profile data, engagement metrics, and thumbnails without API keys.
Featured*
