NSFW JS

What it does

NSFWJS is a JavaScript library for developers adding image-content classification to an application. It accepts browser image, video, or canvas elements and tensors, and can also be used in Node.js with TensorFlow.js. Its model returns category names and probability scores for five classes: Drawing, Hentai, Neutral, Porn, and Sexy. The browser demo lets visitors try classification on their device; the library itself is installed into a project rather than used as a hosted API.

The documented workflow is to install NSFWJS and its TensorFlow.js peer dependency, load a built-in or hosted model, then call classify on image input and consume the returned predictions. Classification can run in a browser or Node.js; React Native is also referenced through a separate mobile demo. Browser classification happens on-device according to the official page, while a Node.js deployment runs in the developer's environment. Results are not definitive: the project says the model is not perfect. Model loading, backend choice, and deployment setup require developer configuration.

What it can do

• Classify images into five content categories
• Return category probability predictions
• Accept browser image, video, and canvas elements
• Load built-in or developer-hosted models
• Select TensorFlow.js execution backends
• Run classification in browser or Node.js

Works with

• TensorFlow.js
• React Native (via separate mobile demo)

Before you choose

• The project states that classification is not perfect; predictions should not be treated as definitive moderation decisions.
• The library is a developer dependency, not a hosted classification API; Node.js deployments may require `@tensorflow/tfjs-node`.
• Configure the model and execution backend for the application. The repository warns that its previous CloudFront model location moved and recommends hosting the model files yourself.

Listed under

• Security
• Client-side
• Content moderation
• Developers
• Image classification
• Javascript
• Nsfw detection

Get started

The site gives `npm install nsfwjs @tensorflow/tfjs` as the installation command.

1. Install `nsfwjs` and the TensorFlow.js peer dependency with `npm install nsfwjs @tensorflow/tfjs`.
2. Load a built-in model, or provide a model URL/path if hosting model files yourself.
3. Pass an image input to `model.classify()` and use the returned class names and probabilities.

No product subscription or usage pricing is stated. Running TensorFlow.js or hosting model files may have separate infrastructure costs; no such costs are specified by the product source.

Source: nsfwjs.com, checked 2 Nov 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 2 Nov 2026. Each source supports part of the overview; availability and prices can change.

• NSFWJS – Client-side NSFW Image Detection in JavaScript ↗nsfwjs.com
• GitHub - infinitered/nsfwjs: NSFW detection on the client-side via TensorFlow.js · GitHub ↗github.com
• NSFWJS – Client-side NSFW Image Detection in JavaScript ↗nsfwjs.com

Company website

nsfwjs.com: product, pricing and policy pages

YC directory mirror

Batch and company name

Human review

Checked by a person

knowledge_base

manual_review

Last checked 2 Nov 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.