WebNet Builder

Design, train and evaluate neural networks without writing code

A no-code platform for building neural networks visually. Import a CSV dataset, assemble an architecture layer by layer, then train and evaluate it directly in the browser — nothing is sent to a server, because the whole training loop runs on your own hardware through TensorFlow.js.

• ▸Visual layer editor for Dense, Dropout, Convolutional and LSTM architectures
• ▸CSV import with MinMax, Standard, Label and One-Hot preprocessing
• ▸In-browser training on WebGPU, WebGL or WebAssembly with live accuracy and loss
• ▸Projects and trained models export as ZIP for sharing or later work

• Angular
• TensorFlow.js
• WebGPU
• TypeScript
• D3.js