Local AI
A real machine-learning model running in your browser — classify sentiment, sort text into labels you invent, or summarize a passage, all locally.
Modern transformer models are usually something you call over an API, sending your text to someone else's server. transformers.js runs them in the browser instead, on the ONNX runtime, so the inference happens on your own machine. Pick a task — sentiment, zero-shot topic classification where you make up the labels, or summarization — type your text, and the model does the rest.
This is the one tool here that fetches its engine from a CDN, because a useful model is tens of megabytes and cannot be bundled the way the smaller WebAssembly tools are. It downloads once and the browser caches it; after that it works offline. The trade-off is deliberate and narrow: the open-source model is fetched, but the text you feed it never leaves the page.
What people open it for
- Gauging the sentiment of a sentence or a review
- Sorting text into categories you define on the spot
- Summarizing a few paragraphs down to the gist
- Seeing a transformer model run entirely client-side