Web UI for AutoGen (A Framework Multi-Agent LLM Applications)
Experimental UI for working with AutoGen agents, based on the AutoGen library. The UI is built using Next.js and web apis built using FastApi.
AutoGen is a framework that enables the development of LLM applications using multiple agents that can converse with each other to solve complex tasks. A UI can help in the development of such applications by enabling rapid prototyping and testing and debugging of agents/agent flows (defining, composing etc) inspecting agent behaviors, and agent outcomes.
Note: This is early work in progress.
Note that you will have to setup your OPENAI_API_KEY or general llm config using an environment variable. Also See this article for how Autogen supports multiple llm providers
export OPENAI_API_KEY=<your key>
Install dependencies. Python 3.9+ is required. You can install from pypi using pip.
pip install autogenui .
or to install from source
git clone [email protected]:victordibia/autogen-ui.git
cd autogenui
pip install -e .
Run ui server.
Set env vars OPENAI_API_KEY
and NEXT_PUBLIC_API_SERVER
.
export OPENAI_API_KEY=<your_key>
autogenui # or with --port 8081
Open http://localhost:8081 in your browser.
To modify the source files, make changes in the frontend source files and run npm run build
to rebuild the frontend.
autogenui --reload
note: the UI loaded by this CLI in a pre-complied version by running the frontend build command show blow. That means if you make changes the frontend code or change the hostname or port the backend is running on the frontend updated frontend code needs to be rebuilt for it to load through this command.
cd frontend
Install dependencies
yarn install
Run in dev mode - with hot-reload
export NEXT_PUBLIC_API_SERVER=http://<your_backend_hostname>/api
your_backend_hostname - is the hostname that autogenui is running on e.g. localhost:8081
yarn dev
(Re)build
Remember to install dependencies and set NEXT_PUBLIC_API_SERVER
before building.
yarn build
@inproceedings{wu2023autogen,
title={AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation Framework},
author={Qingyun Wu and Gagan Bansal and Jieyu Zhang and Yiran Wu and Shaokun Zhang and Erkang Zhu and Beibin Li and Li Jiang and Xiaoyun Zhang and Chi Wang},
year={2023},
eprint={2308.08155},
archivePrefix={arXiv},
primaryClass={cs.AI}
}