fix: garbled voice on tool calls, sync safety, edit-not-duplicate, calmer persona

feat: conversation minimize, HF sync UX, inventory editing/sort/drag/tags

fix: HF sync save — don't fail on 'repo already exists' + surface real errors

feat: vision via native gpt-realtime image input (no extra key scope)

docs: Settings hint — key needs Realtime + Model capabilities permissions

fix: surface the real vision-request error + camera exposure warm-up

feat: real vision — take_picture tool + photo in the conversation

fix: user transcription model + brand icons + date-input overflow

feat: brand-styled app + GA transcription fix + smoothness pass

feat: live conversation transcript panel during voice

feat: robot-only voice — remove phone mode, prime robot mic, visible motion

fix: smooth head-only motion, no audio duplication, mic diagnostics

fix: send TTS to robot speaker via replaceTrack (SDK 1.8 audio path)

fix: reliable robot-mic capture + restore talking motion (daemon wobbler)

fix: disable daemon head-wobbler from app; restore robot-mic capture

fix: stop client-side head-wobble fighting the daemon's native wobbler

fix: smooth speaking head-wobble + persist API key on paste

fix: keep robot audio link muted until a voice session is active

chore: retire Mac-only daemon hacks; point users to Reachy Mini Control

feat: voice — talk to Reachy via OpenAI Realtime, wired to the kitchen tools

feat: onboarding journey — prerequisites on landing + first-run setup

feat: restore the original hand-made landing design + wire app launch

feat: informational landing page + clear new-user journey

feat: use Reachy Mini avatar as app icon (replaces drawn SVG)

build: serve the web app from the repo root (free static Space path)

build: commit web/dist (HF static builder requires account credits)

feat: add mobile-first JS web app (Pollen's new Reachy Mini JS workflow)