Every meaning gets its own pictures

Open any card in Riff and one meaning arrives several ways at once: pictures made for that exact sense, definitions graded to your level, example sentences written for that meaning, and audio for all of it.

That redundancy is deliberate, and it rests on the most tested idea in multimedia education. Richard Mayer’s multimedia principle, backed by decades of controlled experiments, holds that people learn better from words and pictures together than from words alone, provided the two arrive side by side and the clutter stays out. The idea has now been tested on flashcards specifically: a 2025 study comparing multimodal cards carrying text, image and audio against text-only cards found the multimodal ones led to clearly better recall on delayed tests, weeks later, where it actually counts.

The same literature carries a warning that shaped the design as much as the headline finding: extras only help when they belong to the meaning, otherwise they become noise. That is why the images on a card are generated for that one sense rather than pulled from stock, and why a card shows nothing it does not teach.