Metadata-Version: 2.4
Name: rapidllm
Version: 0.4.0
Summary: A small framework for building predictable multi-tool LLM agents.
Author-email: Patrick Tuyizere <you@example.com>
License: MIT
Project-URL: Homepage, https://github.com/Patk-Tuy/rapidllm
Project-URL: Repository, https://github.com/Patk-Tuy/rapidllm
Requires-Python: >=3.8
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: litellm>=0.1.0
Dynamic: license-file

# RapidLLM

RapidLLM is a Python framework for building predictable, reliable multi-tool LLM agents. It emphasizes **multi-tool orchestration** over single-call monolithic prompts, making your agents more modular, testable, and efficient.

## Features
- Simple multi-tool agent architecture.
- Clear separation of concerns between tools.
- Predictable step-by-step execution.
- Easy to extend with your own tools.

## Installation
```bash
pip install rapidllm

## Quick Example

Below is a very minimal example of using RapidLLM to create an agent that lists files:

from rapidllm.agent import run_agent
from rapidllm.tools import list_files_tool

# Define available tools
TOOLS = [list_files_tool]

if __name__ == "__main__":
    run_agent(
        system_prompt="List all files in the src directory.",
        tools=TOOLS,
        model="gpt-4"
    )
