Agentok Studio

AutoGen Visualized - Build Multi-Agent Apps with Drag-and-Drop Simplicity.

⚠️ Warning: This project is currently under development and is not yet recommended for production use.

While code generation has proven effective in maintaining the tool’s utility across various runtimes beyond AutoGen, it has also revealed challenges related to efficiency and observability. As a result, I’m considering transitioning the project into a workflow engine rather than continuing as a standalone program. This change would introduce significant modifications to the backend architecture.

Your feedback is invaluable as we move forward. If you have any thoughts or suggestions, please share them by opening a new issue or participating in the ongoing discussions here.

🌟 What is Agentok Studio

Agentok Studio is a tool built upon AutoGen, a powerful agent framework from Microsoft and a vibrant community of contributors.

Visualizing AutoGen

We consider AutoGen to be at the forefront of next-generation Multi-Agent Applications technology. Agentok Studio takes this concept to the next level by offering intuitive visual tools that streamline the creation and management of complex agent-based workflows. This simplifies the entire process for creators and developers.

Conversation Relations

The relationship between two agents is essential. To incorporate tool calls in a conversation, the LLM must determine which tools to invoke, while informing the user proxy about which nodes to execute. Configuring tools on the edge between these nodes is crucial for optimal operation.

Code Generation

We strive to create a user-friendly tool that generates native Python code with minimal dependencies. Simply put, Agentok Studio is a diagram-based code generator for autogen. The generated code is self-contained and can be executed anywhere as a normal Python program, relying solely on the official pyautogen library.

We also integrated a basic RAG content management feature. This feature is still under development.

Contributions (Issues, Pull Requests, Documentation, even Typo-corrections) to this project are welcome! All contributors will be added to the Contribution Wall.

💡 Quickstart

To quickly explore the features of Agentok Studio, visit https://studio.agentok.ai. While we offer an online deployment of this project, please note that it is not intended for production use. The service level agreement is not guaranteed, and stored data may be wiped due to breaking changes.

After login as Guest or with your OAuth2 account, you can click the Create New Project button to create a new project. The new project comes with a sample workflow. You can click the robot icon flashing on the right bottom to start the conversation.

Due to the limitations of GPT-4 and AutoGen, this simple workflow may not work as expected, but it's a good starting point to understand the basic concepts of Agentic App and Agentok Studio.

For a more in-depth look at the project, please refer to Getting Started.

🐳 Run on Local (with Docker)

The project contains Frontend (Built with Next.js) and Backend service (Built with FastAPI in Python), and have been fully dockerized.

Before running the project, you need to create a .env file in the ui abd api directory and set environment variables.

cp frontend/.env.sample frontend/.env
cp api/.env.sample api/.env
cp api/OAI_CONFIG_LIST.sample api/OAI_CONFIG_LIST

Please be aware that Supabase provides both anon key and service_role key for each project. Please be sure to set anon key to NEXT_PUBLIC_SUPABASE_ANON_KEY for frontend, and service role key to SUPABASE_SERVICE_KEY for backend(api).

The easiest way to run on local is using docker-compose:

docker-compose up -d

You can also build and run the ui and service separately with docker:

docker build -t agentok-api ./api
docker run -d -p 5004:5004 agentok-api

docker build -t agentok-frontend ./frontend
docker run -d -p 2855:2855 agentok-frontend

(The default port number 2855 is the address of our first office.)

🛠️ Run on Local (Without Docker)

If you're interested in contributing to the development of this project or wish to run it from the source code, you have the option to run the ui and service independently. Here's how you can do that:

Frontend

Navigate to the ui directory cd frontend.
Rename .env.sample to .env.local and set the value of variables correctly.
Install the necessary dependencies using the appropriate package manager command (e.g., pnpm install or yarn).
Run the ui service using the start-up script provided (e.g., pnpm dev or yarn dev).

If you see Server Error related to 'useContext' quite often, it's possibly caused by the bugs in turbo mode. In this case, please remove --turbo from the dev command in package.json.

Backend Services

Switch to the api service directory cd api.
Rename .env.sample to .env, OAI_CONFIG_LIST.sample to OAI_CONFIG_LIST, and set the value of variables correctly.
Install Poetry.
Launch with command poetry run uvicorn agentok_api.main:app --reload --port 5004.

REPLICATE_API_TOKEN is needed for LLaVa agent. If you need to use this agent, make sure to include this token in environment variables.

IMPORTANT: The latest version of AutoGen requires Docker for code execution by default. To proceed, you must either:

Install Docker on your local machine, OR
Disable this requirement by setting AUTOGEN_USE_DOCKER=False in the api/.env file.

Note: This requirement is disabled by default since the default deployment of this project is already dockerized.

Database Services

This project relies on Supabase for user authentication and data storage. To get started, please follow the ./db/README.md to prepare the database, and set the environment variables (Refer to those variables with name starts with SUPABSE in .env.sample) in the .env file.

If you prefer, you can deploy your own Supabase instance, but that is beyond the scope of this document.

Once you've started both the frontend and api services by following the steps previously outlined, you can access the application by opening your web browser and navigating to:

api: http://localhost:5004 (OpenAPI docs served at http://localhost:5004/docs)
frontend: http://localhost:2855

If your services are started successfully and running on the expected ports, you should see the user interface or receive responses from the api services via this URL.

👨‍💻 Contributing

Contributions are welcome! It's not limited to code, but also includes documentation and other aspects of the project. You can open a GitHub Issue or leave comments on our Discord Server.

This project welcomes contributions and suggestions. Please read our Contributing Guide first.

If you are new to GitHub, here is a detailed help source on getting involved with development on GitHub.

Please consider contributing to AutoGen, as Agentok Studio relies on a robust foundation to deliver its capabilities. Your contributions can help enhance the platform's core functionalities, ensuring a more seamless and efficient development experience for Multi-Agent Applications.

This project uses 📦🚀semantic-release to manage versioning and releases. To avoid too frequent auto-releases, we make it a manual GitHub Action to trigger the release.

To follow the Semantic Release process, we enforced commit-lint convention on commit messages. Please refer to Commitlint for more details.

Contributors Wall

📝 License

The project is licensed under Apache 2.0 with additional terms and conditions.

lyu-ai / agentok