This project demonstrates how to extend the agent-framework by creating a custom chat client for Google's Gemini large language models. It includes a working GeminiChatClient that implements the ChatClientProtocol and an example script showing how to use it.

• Custom Chat Client: Implements GeminiChatClient to connect to the Google Gemini API.
• Protocol Compliant: The client correctly implements the ChatClientProtocol.
• Streaming and Non-Streaming: Supports both standard request-response and streaming responses.
• Easy to Use: A simple main.py script demonstrates how to instantiate and run an agent with the custom client.

• Python 3.8+
• A Google Gemini API Key. You can get one from Google AI Studio.

1. Clone the repository or download the files.

2. Install the required dependencies:

   pip install -r requirements.txt

   or

   uv sync

3. Create a .env file:

   In the root directory of the project, create a file named .env.

4. Add your API Key to the .env file:

   Open the .env file and add your Google Gemini API key like this:

   GEMINI_API_KEY="YOUR_API_KEY_HERE"
   

Once the setup is complete, you can run the example script:

python main.py

or

uv run main.py

The script will execute two interactions with the Gemini model:

1. A non-streaming request for the "meaning of life," printing the complete JSON response at the end.
2. A streaming request for the "purpose of life," printing the response text to the console as it is generated.

The output will look something like this:

Agent response: {"messages": [{"role": "assistant", "contents": [{"text": "The meaning of life is a profound question that has been debated by philosophers, theologians, and thinkers for centuries..."}]}], "response_id": "..."}
Agent streaming response:
The purpose of life, much like its meaning, is a deeply personal and philosophical question...

├── main.py                 # Main entry point, demonstrates agent usage.
├── gemini.py               # Contains the GeminiChatClient implementation.
├── requirements.txt        # Project dependencies.
├── .env                    # (You create this) For storing environment variables.
└── README.md               # This file.

• main.py: This script shows how to instantiate ChatAgent with the custom GeminiChatClient and run it in both non-streaming (agent.run) and streaming (agent.run_stream) modes.
• gemini.py: This module contains the GeminiChatClient class. It handles the logic for communicating with the Google Gemini API, including message formatting and handling both single and streaming responses.