raw-agents 0.1.0

A powerful and flexible LLM interface library for Gemini and other models.

RAW Library

A powerful and flexible LLM interface library for Gemini and other models, designed for modern Python applications.

Features

• Gemini Integration: Seamless support for Google's Gemini models (including gemini-3-flash-preview).
• Unified Interface: Clean, consistent API for chat, completion, and tool usage.
• Async Support: Fully asynchronous implementation using httpx and asyncio.
• Tool Calling: Easy-to-use function calling/tool usage.
• Streaming: Native support for streaming responses.
• Vision Support: Simple API for multimodal (text + image) inputs.

Installation

pip install raw

or with uv:

uv add raw

Quick Start

Configuration

Set your Gemini API key as an environment variable:

export GEMINI_API_KEY="your-api-key-here"

Basic Usage

import asyncio
import os
from RAW.llms.gemini import GeminiLLM

async def main():
    api_key = os.getenv("GEMINI_API_KEY")
    llm = GeminiLLM(api_key=api_key)
    
    response = await llm.generate("Explain quantum computing in one sentence.")
    print(response)

asyncio.run(main())

Chat with Tools

from RAW.llms.gemini import GeminiLLM
from RAW.modals import Message, Tool, ToolParam

# Define a tool
def get_weather(city: str):
    return f"The weather in {city} is sunny."

tool = Tool(
    name="get_weather",
    description="Get weather for a city",
    parameters=[ToolParam(name="city", type="string", required=True)],
    function=get_weather
)

# Chat with tool
messages = [Message(role="user", content="What's the weather in Tokyo?")]
response = await llm.chat(messages, tools=[tool])

License

MIT