brainlabz / ai-flow

Open Source user-friendly UI application that allows you to easily connect multiple AI models together.

Open-source tool to seamlessly connect multiple AI models in repeatable flow.

🔗 Website

While a live demo is available for convenience, for the best experience we recommend running the application directly on your local machine.

🎉🚀 v0.5.1 is Now Available 🚀🎉

🚀 GPT-4-Turbo | 👀 GPT-4-Vision | 🎨 Dall-E-3
🌟 UI Enhancements | ↔️ New Layout View | 🤖 New Merge Node | ✨ Support for large input

---

AI Flow is an open source, user-friendly UI application that empowers you to seamlessly connect multiple AI models together, specifically leveraging the capabilities of ChatGPT. This unique tool paves the way to creating interactive networks of different AI models, fostering a multi-perspective response to various prompts.

Imagine setting up a simple workflow where:

• One node generates a poem
• One node gives a review of the poem
• One node creates a prompt for the cover of the poem
• One node makes the cover with Stable Diffusion

If we want to change the idea of the poem, we can edit the first node, and keep the same flow.

This flow logic can then be applied to different inputs, yielding varied results for similar subject matter.

AI Flow is particularly useful for content creation. By enabling different GPT models to adopt different roles, AI Flow orchestrates the creation and evaluation of digital content, from social media threads to blog articles.

With AI Flow, you utilize your own OpenAI keys, giving you full authority over your budget. Importantly, the software runs locally on your machine, assuring you of privacy and control over your data and operations.

⚠️ Please note that very large input (for example the transcript of a 1 hour podcast) may incur additionnal costs, because for this type of entry the default behavior is to split the input into chunks, create embeddings (with the OpenAI API), and then use a similarity search to find the right items to answer the prompt.

N.B : For large inputs, you'll get better result with a precise prompt.

Features

In a nutshell, AI Flow provides a visual platform for crafting and managing AI-driven workflows, thereby facilitating diverse and dynamic AI interactions.

• 🎨 It offers a drag-and-drop interface to design these workflows
• 📊 Monitors their execution in real-time
• 🗂️ AI models can be conveniently managed and organized
• 💾 Workflows can be exported or imported for sharing or backup purposes

Other basic use cases

Summarize web pages & youtube transcript

Use pre-made prompts to build quickly

Split input with AI

By combining those features we can create different flows.

Describe picture with AI

Create a custom layout for your flow

When your flow is complete, you have the option to run it as is or create a custom layout focusing on the most interesting parts.

Installation (Windows executable)

For a quick local setup, grab the .exe from the repository's releases section.

Installation (without Docker)

Prerequisites

Before getting started, make sure you have the following dependencies installed on your system:

• Python (version 3.9.5 or later)
• Poetry (version 1.4.2 or later)
• Node.js (version 16.13.0 or later)

Clone the Repository

1. Clone the repository: git clone https://github.com/DahnM20/ai-flow.git
2. Change to the project directory: cd ai-flow

UI Dependencies

1. Go to the UI directory: cd packages/ui
2. Install dependencies: npm install

Backend Dependencies

1. Go to the backend directory: cd packages/backend
2. Install Python dependencies: poetry install

For Windows only

3. Launch poetry shell : poetry shell
4. Install the windows requirements in the poetry shell : pip install -r requirements_windows.txt

Usage

1. Start the backend server: cd backend && poetry run python server.py
2. Start the ui application: cd ui && npm start
3. Open your browser and navigate to http://localhost:3000
4. Use the drag-and-drop interface to design your AI workflow
5. Connect AI models and define data flow between them
6. Click "Run" to execute the AI workflow
7. Monitor the execution progress and results in real-time

🐳 Docker

Docker Compose

1. Go to the docker directory: cd ./docker
2. Update the .yml if needed for the PORTS
3. Launch docker-compose up or docker-compose up -d
4. Open your browser and navigate to http://localhost:80
5. Use docker-compose stop when you want to stop the app.

Contributing

Contributions are welcome! If you find any issues or have suggestions for improvement, please open an issue or submit a pull request.

License

This project is licensed under the MIT License. See the LICENSE file for more details.