Google DeepMind Models Samples, Snippets and Guides

This repository contains personal tiny samples, snippets and guides showcasing cool experiments and implementations using Google DeepMind Gemini models.

├── examples/
│   └── gemini-with-openai-sdk.ipynb
├── guides/
│   └── function-calling.ipynb
├── scripts/
│   └── gemini-browser-use.py
├── javascript-examples/
│   └── gemini-native-image-out.js
├── LICENSE
└── README.md

Guides:

• Function Calling Guide - Comprehensive guide on implementing function calling
• ReAct Agent - Guide on building a ReAct agent with LangGraph and Gemini
• Agentic Patterns - Overview of agentic patterns and how to use them with Gemini
• Gemini with Memory - Example how to integrate long-term memory into Gemini

Examples

• Gemini with OpenAI SDK - Use Gemini models with the OpenAI SDK
• Gemini with Google Search - Enable Google Search integration for up-to-date information
• Structured Outputs - Generate structured JSON outputs using Pydantic
• Meta Prompts - Generate JSON schemas dynamically using meta prompts
• Gemini MCP Example - Example showcasing Model Context Protocol (MCP)
• Gemini Analyze & Transcribe YouTube - Analyze and transcribe YouTube videos
• Gemini Native Image Output - Generate images directly with Gemini 2.0 Flash experimental model
• Gemini File Editing - Example showcasing file editing capabilities
• Gemini LangChain Integration - Use Gemini models with LangChain
• Gemini Code Executor Data Analysis - Example of data analysis using Gemini code execution
• Gemini ADK MCP - Example using Gemini with ADK for MCP
• Gemini PydanticAI Agent - Build agents using Gemini and PydanticAI
• Gemini Context Caching - Example showcasing context caching and saving up to 75% on costs with Gemini
• Gemini CrewAI - Example showcasing CrewAI with Gemini 2.5 Pro Preview
• Sequential Function Calling - Example showcasing sequential function calling with Gemini
• Gemini Batch API - Example showcasing batch API with Gemini
• Gemini Context URL - Example on how to use URLs as context with Gemini
• Gemini Few-shot PDF - Few-shot PDF extraction with Gemini
• Gemini Text-to-Speech - Example for using text-to-speech with Gemini
• OpenAI SDK with Gemini - Use OpenAI SDK with Gemini

Scripts

• Gemini Browser Use - Example script for using Gemini with browser interaction.
• Gemini MCP Agent - A basic agent script demonstrating MCP.
• Gemini MCP Pipedream - A basic agent script demonstrating MCP with Pipedream.
• Gemini Veo3 Optimized Prompt - A script using Gemini 2.5 to optimize prompts for Gemini Veo3.
• Gemini Veo3 Vlogs - Automatic video generation with Gemini Veo3 combined multiple video clips.
• Gemini Image Meta - Generate images with Gemini and Imagen.

JavaScript Examples

• Gemini Native Image Output - Generate images directly with Gemini 2.0 Flash experimental model

Gemma

• Gemma with GenAI SDK - Use Gemma 3 27B It with Google's GenAI API
• Gemma Function Calling - Implement function calling with Gemma 3 27B

How to Use

1. Clone the repository:

   git clone https://github.com/philschmid/gemini-samples.git
   

2. Set up environment variables: Create a .env file in the root directory:

   GEMINI_API_KEY=your_api_key_here
   

3. Explore the examples: Browse the sample notebooks to find code related to different DeepMind models and experiments.

4. Run and modify: Experiment with the code, tweak parameters, and integrate the snippets into your own projects.

Contributing

Contributions are welcome! If you have additional examples or improvements, please feel free to open a pull request or create an issue.

License

This repository is licensed under the MIT License. See the LICENSE file for more details.