A visual node-based workflow editor for configuring data processing pipelines. Built with vanilla JavaScript, Tailwind CSS, Flask, and SQLite.
- 🎨 Modern Dark UI - Professional OLED dark theme optimized for developer tools
- 🔌 Drag & Drop - Intuitive drag-and-drop interface for building workflows
- 🔗 Visual Connections - Smooth bezier curves connecting node inputs and outputs
- ⚙️ Live Parameter Editing - Edit node parameters in real-time with type-aware inputs
- 💾 SQLite Backend - All component data stored in SQLite, easy to extend
- 📱 Responsive Design - Works on desktop, tablet, and mobile
- ⌨️ Keyboard Shortcuts - Full keyboard navigation support
- 💨 Auto-save - Automatic persistence to localStorage
- Drag components from sidebar to canvas
- Connect nodes by dragging between ports
- Edit parameters in the properties panel
- Pan and zoom the canvas
- Delete nodes and connections
- Float - Decimal numbers with step control
- Integer - Whole numbers
- Boolean - True/False radio buttons
- String - Text input or dropdown (with options)
pip install -r requirements.txtpython3 api.pyThis will:
- Create
components.dbfrom the schema - Start Flask API on http://localhost:5000
Option A - Direct file:
open node-editor.htmlOption B - Local server:
python3 -m http.server 9999
# Then open: http://localhost:9999/node-editor.htmlautora2/
├── node-editor.html # Main frontend application
├── api.py # Flask REST API server
├── components.db.sql # Database schema and sample data
├── components.db # SQLite database (auto-generated)
├── requirements.txt # Python dependencies
├── setup.sh # Setup script
├── NODE_EDITOR_DESIGN.md # Complete design specification
├── NODE_EDITOR_README.md # User guide
└── README_SQLITE.md # SQLite backend documentation
┌─────────────────┐ HTTP ┌─────────────────┐ SQL ┌─────────────────┐
│ node-editor │ ──────────────> │ Flask API │ ────────────> │ SQLite DB │
│ .html │ GET /api/ │ (api.py) │ SELECT * │ (components │
│ (Frontend) │ components │ (Backend) │ FROM ... │ .db) │
└─────────────────┘ <────────────── └─────────────────┘ <──────────── └─────────────────┘
JSON Results
- Drag a component from the left sidebar (e.g.,
darts.json) - Drop it anywhere on the canvas
- Repeat to add more nodes
- Click and hold the output port (right side) of a node
- Drag to the input port (left side) of another node
- Release to create the connection
- Valid ports highlight green when hovering
- Click a node to select it
- The properties panel slides in from the right
- Modify parameter values
- Click "Save Changes" to apply
- Pan: Click and drag on empty canvas
- Zoom: Mouse wheel or +/- buttons
- Fit: Click fit-to-screen button
| Shortcut | Action |
|---|---|
Delete / Backspace |
Delete selected node |
Escape |
Close properties panel |
Ctrl/Cmd + S |
Save canvas |
Returns all components with their parameters.
Response:
[
{
"name": "darts",
"parameters": [
{
"name": "x",
"datatype": "float",
"defaultValue": 0.0,
"description": "X coordinate"
}
]
}
]CREATE TABLE components (
id INTEGER PRIMARY KEY AUTOINCREMENT,
name TEXT NOT NULL UNIQUE,
created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
);CREATE TABLE parameters (
id INTEGER PRIMARY KEY AUTOINCREMENT,
component_id INTEGER NOT NULL,
name TEXT NOT NULL,
datatype TEXT NOT NULL,
default_value TEXT,
description TEXT,
options TEXT,
FOREIGN KEY (component_id) REFERENCES components(id)
);sqlite3 components.dbINSERT INTO components (name) VALUES ('my_component');
INSERT INTO parameters (component_id, name, datatype, default_value, description)
VALUES (6, 'my_param', 'float', '1.0', 'My parameter description');import sqlite3
conn = sqlite3.connect('components.db')
cursor = conn.cursor()
cursor.execute("INSERT INTO components (name) VALUES (?)", ('my_component',))
component_id = cursor.lastrowid
cursor.execute("""
INSERT INTO parameters (component_id, name, datatype, default_value, description)
VALUES (?, ?, ?, ?, ?)
""", (component_id, 'my_param', 'float', '1.0', 'Description'))
conn.commit()
conn.close()Edit components.db.sql, then:
rm components.db
python3 api.py # Recreates from schemaThe editor includes 5 pre-configured components:
- darts - Coordinate-based processing (2 parameters)
- linear_model - ML model configuration (4 parameters)
- data_processor - Data pipeline settings (3 parameters)
- evaluator - Model evaluation metrics (2 parameters)
- neural_network - Neural network architecture (3 parameters)
- Frontend: Vanilla JavaScript, Tailwind CSS, HTML5
- Backend: Flask 3.0, Python 3.8+
- Database: SQLite 3
- Icons: SVG (Heroicons style)
- Fonts: IBM Plex Sans, JetBrains Mono (Google Fonts)
- Chrome 90+
- Firefox 88+
- Safari 14+
- Edge 90+
- KISS - Keep it simple, no over-engineering
- No unnecessary features - Only what's needed
- Simple code structure - Easy to understand and modify
- Minimal dependencies - Vanilla JS, no heavy frameworks
sqlite3 components.db
sqlite> .tables
sqlite> SELECT * FROM components;rm components.db
python3 api.pycurl http://localhost:5000/api/components | jq- Design Specification - Complete UI/UX design guide
- User Guide - Detailed usage instructions
- SQLite Backend - Database architecture and API docs
- Ensure Flask API is running:
python3 api.py - Check API is accessible:
curl http://localhost:5000/api/components
flask-corsshould handle this automatically- Check browser console for details
- Only one process can write to SQLite at a time
- Close other connections to
components.db
- Undo/redo functionality
- Multi-node selection
- Connection type validation
- Grid snapping toggle
- Minimap for large canvases
- Node grouping
- Copy/paste nodes
- Import/export JSON workflows
- Export to image
- Touch gestures for mobile
MIT License - feel free to use and modify
Contributions welcome! Please:
- Fork the repository
- Create a feature branch
- Make your changes
- Submit a pull request
Created with Claude Code
Version: 1.0.0 Last Updated: 2026-01-16