Multilingual Sentiment Analysis for Selected Languages in Africa and Beyond

This application provides multilingual sentiment analysis using a Glot500-based model. It enables sentiment prediction for text in multiple languages, including selected languages spoken in Africa and other multilingual contexts. The model aims to support NLP research for underrepresented and low-resource languages. Inference runs entirely in your browser — no server, no data leaves your device.

On language coverage: the model is multilingual and inherits broad language coverage from the Glot500 pretraining framework (~500 languages). The fine-tuning dataset does not contain explicit language annotations, so the exact language distribution cannot be fully determined. Script-level analysis and manual inspection indicate the presence of multiple languages, including selected languages spoken in Africa and other multilingual contexts (see the README for details). This is not an exhaustive list of supported languages, and results for languages/scripts not well represented in fine-tuning (e.g. Tigrinya) may be less reliable.

Input Text

Loading model… this can take a while on first visit (~630 MB, cached afterward).

Prediction—

Confidence—

Raw model label—