Project Demo Video
Experimental Visual Machine Learning IDE built with C++ & Qt
tangleML is a visual IDE designed to simplify machine learning workflows by showing how information moves through connected nodes and reducing repetitive code. Inspired by Blender’s node system, it allows users to drag boxes onto a canvas, connect them, and generate Python code automatically.
Underlying engine status:
An independent execution engine was planned as part of the architecture, but development is still incomplete. The current experimental engine repository is available here: TangleML Engine
-
Drag-and-drop workflow: Add boxes from the toolbox to the canvas and connect pointers to define data flow.
-
Canvas → Python file sync: Saving a project updates the Python file with the code represented by the boxes. (Ctrl+S)
-
Run small code snippets: Test code in the Python file without affecting the canvas setup.
-
Console with dual output:
- Upper area: standard print/log output.
- Lower area: images and graphs (e.g., from matplotlib).
-
Reusable custom boxes: Create permanent custom nodes via the menu bar (
File → Add File) for code that can be reused across projects. -
Temporary inline code box: Drag the custom box onto the canvas, (
double-click it), and write temporary code directly inside the node. -
Rapid ML prototyping: Quickly create workflows without rewriting boilerplate ML code.
tangleML is experimental and incomplete:
- Projects cannot be reopened in tangleML once closed (Python files remain accessible).
- Most menu bar functions are placeholders.
- Toolbox templates for common ML tasks are not included yet.
- No logo, branding, or polished UI.
- Only basic box connections and Python code generation are functional.
Development of tangleML was supported by ChatGPT, which assisted with debugging, code structure, and logic design throughout the project.
-
Clone the repository:
git clone https://github.com/6Asahi9/Tangle-UI.git
-
Open the project in Qt Creator or Vscode.
-
Then build and Run
cd/ build,TangleML-UI.
- Drag boxes from the toolbox to the canvas.
- Connect pointers to define data flow.
- Write small code snippets in the Python file tab and run them.
- Save (
Ctrl+S) to sync canvas boxes to the Python file. - Use custom boxes for reusable code snippets via the menu bar.
- Support for saving and loading projects.
- Predefined ML toolbox templates.
- Node property editor for enhanced customization.
- Polished UI/UX with logo and themes.
Contributions are welcome! Since this is an experimental IDE, please:
- Report bugs or limitations in the Issues section.
- Suggest new features or improvements.
- Submit pull requests with clear descriptions of changes.
This project is experimental and intended for learning and prototyping purposes. Not for commercial use.