What's New • New "Discover" gallery — play featured mini-games, personality tests and handy tools for free, no credits needed. • Home now suggests ideas: say one line and make your own version. • Sturdier voice connection — reconnects instantly when your network switches or comes back. • Better generation — upgraded model and more reliable game gestures.

• Tasks no longer break when you switch apps, lock the screen, or background Meco — they keep running and the result waits for you. • Re-open Meco after force-quitting and any task from the last 24 hours is picked back up automatically — no duplicate charges. • Cancelling a task now stops the work instantly — no more silent credit drain in the background. • Mini-games, multi-page decks and complex tools now go through an extra AI self-review pass — mobile tap responsiveness, gesture direction and game loops are significantly more reliable.

· AI songwriting: lyrics on the page now match the actual singing word-for-word · Music result pages have more visual variety — no longer all stuck inside the same player shell · Fixed an issue where toggling off silent mode could snap back to silent after a reconnect · Fixed occasional duplicate task creation when voice commands fire in quick succession · Smoother guidance when you actually want a sung song with real lyrics

Added: "Start Live Conversation" prompt button

v1.1.3 Release Notes New Features Instant Voice Execution — Saying "generate for me" now triggers the task immediately, no follow-up needed Enhanced Music Generation — Generate multiple AI-original tracks with different themes (2 variants each to choose from) Result Subscription Refresh — News/report results can auto-refresh on a schedule to stay current Voice Chat Prompt — A "Start Voice Chat" capsule button appears in silent mode for easier discovery Music Capability Guidance — Naturally guides users toward AI-original music when song/music intent is detected Bug Fixes Fix Duplicate Tasks — A single request no longer creates multiple identical tasks Fix Silent Mode Persistence — Unmuting no longer reverts on reconnection Backend Improvements Unified R2 Static Routes — Config-driven routing, easier to extend Result Page Owner Detection — Enables owner-only actions (e.g. retry music generation) Music Prompt Sanitization — Dual-layer protection against copyright-sensitive words causing generation failures

Upgraded core model to Claude Opus 4.7 for warmer, more natural conversations that learn your style on the fly; smoother silent-mode behavior across reconnects; fixed keyboard-input memory and profile-edit layout glitches.

• Upgraded to Claude Opus 4.7 for tougher tasks • Warmer, more natural chat persona — less AI-assistant, more friend • Fixed: keyboard input now properly remembered across sessions • Profile page: smoother layout when editing your name

· Improved haptic feedback with tactile response on button taps · Added silent mode toggle notification at the bottom of the screen · Merged camera and screen sharing into a single menu for simpler access · Updated onboarding tip to guide users to the voice chat entry · Optimized live video / screen share preview window position · Streamlined task status display area for a cleaner interface · Silent mode now enabled by default for work-friendly experience