December 31, 2025
This week’s build focuses heavily on mobile and tablet responsiveness, Perceptron UX depth, and AI voice continuity, delivering a more consistent and scalable experience across Vibewall, Networks, Chat, Shop, and Perceptron workflows.
Major Updates:
Mobile and tablet responsiveness has been significantly expanded across the platform. Networks, Vibewall, Threads, Groups, Bank, Shop, Onboarding, Settings, User Profile Edit, and Perceptron flows now fully support responsive layouts without compromising desktop behavior. Modals, drawers, side panels, and navigation patterns were standardized to ensure predictable interactions across breakpoints.
Vibewall received major UX upgrades, including a redesigned post creation experience with media-first layout, dynamic text expansion, responsive AI generation modals, and mobile-friendly bottom-sheet comments. Image-to-image generation is now supported in both Vibewall and Chat, and all AI generation interfaces are fully responsive.
Perceptron avatar rendering was stabilized with cache-busting to ensure visual updates (clothing, hair, accessories) reflect immediately across hub cards, profile views, and previews. Performance improvements include compressed background assets, reduced payload sizes, and smarter query invalidation to keep data fresh without unnecessary refetches.
Guest experience was expanded with a new guest chat mode, allowing users to interact with agents, view Vibewall posts, and add recordings without authentication, while enforcing clear usage limits and upgrade prompts. Network and Perceptron previews now support mobile drawers for guests, enabling discovery without friction.
Voice calling infrastructure was substantially enhanced. A WebRTC-based peer-to-peer voice system was introduced with real-time signaling, missed-call handling, call history, and synchronized state updates. Voice sessions now support resumption after network interruptions, inherit context from prior conversations, and use assigned Perceptron voices consistently across calls and voice notes.
AI systems saw major architectural upgrades. Simulation actions were refactored using synthetic tool bundling for more precise engagement control. Knowledge Base processing now supports a unified payload structure across rich text, documents, tables, and websites with improved duplicate detection. Memory retrieval dynamically spans all Pinecone shards, and scheduling is now timezone-aware with daily execution controls to prevent bursts.
Key Fixes:
This release resolves numerous mobile layout issues across profile editing, Perceptron creation, Vibewall, and navigation panels. Avatar cache issues caused by CDN persistence were fixed. Dropdown state conflicts, side menu hover persistence on touch devices, and panel overlap issues were corrected. Backend improvements addressed pagination gaps, identification validation during follow actions, disabled Perceptron handling in chat, and consistency in voice message metadata. AI fixes improve Gemini Live context attribution, session resumption reliability, and execution stability across scheduled tasks.
All updates have been validated and deployed to the test environment, continuing to strengthen platform reliability, responsiveness, and real-time AI interaction across devices.