Preview image. Unlock full-res
Dark bento grid highlighting context-aware mentions, personal vocabulary, multilingual voice, premium model picker, and cross-app search.
Summary
Personalized talk-to-text bento with five glass cards covering @mentions, custom vocabulary, 50+ languages, multi-model chat, and unified app search.
Visual description
Black page with wand-icon pill "Personalized Talk to Text," headline about AI with full context, and dictation subcopy. Five rounded glass cards in an asymmetric grid: top wide card shows a chat message with pink @Ellie and linked doc auto-connected; middle left Personal vocabulary word cloud with highlighted IMO button; middle right Global voice with US/Brazil toggle and faded flag icons; bottom left blue icon card with model selector bar for Gemini, Sonnet, ChatGPT, Brain M1, Deepseek; bottom right green search card querying an OOH campaign with Slack, Drive, Figma, GitHub icons and two file results.
Key takeaway
Asymmetric bento lets one hero capability span full width while supporting features tile below. Mini UI scenes inside each card teach features faster than bullet lists.
Reuse notes
Ideal for AI assistant pages with many small capabilities. Card count and sizes need art direction to avoid visual noise on smaller breakpoints.
