Intelligence you can run.
Four working AI tasks. Real model inference on your device, with explicit downloads and a local execution record.
Read the tone of a sentence.
Classify English text as positive or negative.
The selected model downloads only when you choose. Download size and loading time vary. Inputs are processed locally; model downloads use Hugging Face and jsDelivr.
No prompts or answers are stored in the execution history.
Model output
Your deviceLoad the model, then run your task.
Model
Xenova/distilbert-base-uncased-finetuned-sst-2-english
Quantized q8 · WebAssembly. Scores indicate model output, not factual certainty. Review answers against the source. Similarity scores do not prove equivalence.
Model informationThese on-device tasks make no paid provider request and consume no purchased Resource Credits.
Execution history
Your completed model runs appear here. Records stay in this browser.