Simple Agent Final is a simple and efficient agent system that combines a powerful core and many modules/tools that can be flexibly combined or new ones created.
Install requirements
pip install -r requirements.txt
Set up the API keys in .env and load dotenv in the biggining of code:
from dotenv import load_dotenv
load_dotenv(".env")
from core.agent import Agent
ag = Agent(history=['u', 'hi'], model_name="provider_name/model_name")
response = ag.chat()
print(response)
history = [
['u', 'use ls']
]
ag = Agent(history=history,
model_name="provider_name/model_name",
cwd="~",
preset="tools")
print(ag.chat(loop=1)
ag.add("make screenshot")
print(ag.chat(loop=2)
import json
with open('chat.json', 'w', encoding='utf-8') as f:
json.dump(ag.client.history, f, ensure_ascii=False, indent=4)
after that...
python app.py
- Native saving is planned
This project is distributed under a custom license. By using, forking, or modifying this software, you agree to the following mandatory naming condition.
For full details, please read the LICENSE file.