Local-first, offline AI UI primitives + blocks - run ML models (chat, RAG, vision, audio) entirely in the browser. No servers, no API keys.
Language-grouped TTS voice selection from one KokoroVoice[] contract — a compact grouped <select> (VoicePicker) plus rich VoiceCard / VoiceGrid variants with a color-coded gender badge, monospace voice id, a circular play/stop preview button with loading state, a per-group count header, and search/filter. Wire onPreview to useSynthesizeSpeech for local Kokoro samples.
A pure-CSS row of pulsing vertical bars (sinusoid-derived heights, staggered animation) that doubles as an active audio-processing indicator and an idle empty-state. Supports agent-state modes (connecting/listening/thinking/speaking), record/playback-scrub modes, and a live-volume mode that scales amplitude from a callback — decoupling the visual from the audio source. Ships its keyframe inline so it animates standalone after install.
A draft editor with live length feedback, preset and free-text edit instructions with prompt improvement, and an AI edit reviewed as a before/after diff with explicit Accept/Reject (Accept applies exactly the after text). Resolves the Chrome Prompt API when available (zero download) and automatically falls back to a Llama-3.2-1B Transformers.js instruct model behind an explicit download gate, with a truthful provider badge. No model bytes download until an explicit in-block action.