Portfolio

Lugemi Grounded

Speech plus the selected visual referent. Returns document evidence, speaker claim, and translation separately. Highlights the referenced region and supports wrong-region correction. Source/target languages cover the full registry. Assistive document communication only.

Move between models, verified interpretation, and the playground without leaving the Lugemi console. Each pillar below links to its working surface.

Models · Verified Interpreter · Playground

API keyDocument referenceDocument text (selected region OCR stand-in)Spoken / typed utterance

pagexywidthheight

Source language

Target language

Simulate wrong / empty region (request clarification)

Lugemi — Africa-first language intelligence for speaking agents