cazala/synaptic

Health v2 · maintenance only

Flags: no_license

Adoption not part of the score

6912 stars · 651 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Synaptic is an architecture-free neural network library for JavaScript that runs in both Node.js and the browser. It supports building and training arbitrary first- and second-order neural network architectures, with built-in architectures like multilayer perceptrons, LSTM, liquid state machines, and Hopfield networks.

Use cases

• build and train neural networks in JavaScript
• create LSTM networks in the browser
• train a perceptron to solve XOR
• experiment with different neural network architectures
• learn neural networks with JavaScript examples
• implement recurrent neural networks in Node.js

When to choose

• you need neural networks running in the browser or Node.js with no native dependencies
• you want to define custom network architectures rather than fixed ones
• you are learning or teaching neural network fundamentals in JavaScript

When to avoid

• you need GPU acceleration or high-performance training on large datasets
• you need modern deep learning features like convolutions, transformers, or automatic differentiation
• you require an actively developed library with recent updates and strong community support

Facets

library · maturity maintenance

machine-learning deep-learning machine-learning deep-learning browser neural-networks lstm perceptron hopfield-network recurrent-neural-networks browser-compatible javascript nodejs

Member repositories

For agents

markdown · JSON · MCP: product_card(name="cazala/synaptic")