What is recurrentjs?

karpathy/recurrentjs — explained in plain English

Analysis updated 2026-07-18 · repo last pushed 2016-11-04

★ 984HTMLAudience · developerComplexity · 4/5DormantSetup · moderate

Mindmap

mindmap
  root((recurrentjs))
    What it does
      Builds neural networks
      Runs in the browser
      Automatic differentiation
    Tech stack
      JavaScript
      Graph computation
      RNN and LSTM
    Use cases
      Predict next character
      Generate text
      Prototype sequence models
    Audience
      Researchers
      Students
      Web developers
    Tradeoffs
      Beta quality
      Needs ML knowledge

mindmap root((recurrentjs)) What it does Builds neural networks Runs in the browser Automatic differentiation Tech stack JavaScript Graph computation RNN and LSTM Use cases Predict next character Generate text Prototype sequence models Audience Researchers Students Web developers Tradeoffs Beta quality Needs ML knowledge

What do people build with it?

USE CASE 1

Train a character-level text generator that runs entirely in the browser, no backend needed.

USE CASE 2

Prototype a sequence model quickly to demo how RNNs and LSTMs learn from data.

USE CASE 3

Study automatic differentiation and backpropagation by reading and running the library's Graph code live.

USE CASE 4

Add text generation or sequence prediction features to an interactive web project.

What is it built with?

JavaScriptHTML

How does it compare?

	karpathy/recurrentjs	codingdocs/typora-markdown-resume	gustavoguanabara/projeto-site
Stars	984	907	902
Language	HTML	HTML	HTML
Last pushed	2016-11-04	—	2024-07-20
Maintenance	Dormant	—	Stale
Setup difficulty	moderate	easy	easy
Complexity	4/5	1/5	1/5
Audience	developer	developer	vibe coder

Figures from each repo's GitHub metadata at analysis time.

How do you get it running?

Difficulty · moderate Time to first run · 1h+

Beta-quality code, requires understanding neural networks and backpropagation to use effectively.

Copy-paste prompts

Prompt 1

Show me how to build a character-level RNN with recurrentjs and train it on a sample text file.

Prompt 2

Explain how recurrentjs's Graph system performs automatic differentiation through matrix operations.

Prompt 3

Help me set up a loss function and training loop for an LSTM model using recurrentjs.

Prompt 4

Walk me through the demo where recurrentjs predicts the next letter in a sentence.

Prompt 5

Compare using recurrentjs's RNN versus LSTM architecture for a sequence-prediction task.

Frequently asked questions