October 27, 2025
Build v0.1.20
This release focuses on UI refinement, backend efficiency, and AI intelligence upgrades. It consolidates the Search Module, enhances Perceptron functionality with new host tools and task APIs, improves Thread and Whisper experiences, and refines Knowledge Base workflows for more structured AI reasoning. The update strengthens platform performance, scalability, and overall user experience.
Major updates
Added “Get Beta Key” and “Book a Demo” flows for Early Access.
Introduced new Networks wallpapers and Perceptron Sleeptime variations.
Conducted research and design for Settings Module (WIP).
Unified Search Module: merged MainSearchPage and SearchModal into a single structure, removing redundant components and improving maintainability.
Perceptron Host Toolbar:
Integrated Scheduled Tasks API with React Query caching and auto-refresh.
Added new Traces tab with collapsible sections, search, pagination, and detailed output panels.
Improved visual consistency with shimmer loaders and better overflow handling.
Perceptron Memory Analytics: Integrated V2 BarChart with live API data.
Notifications: Improved loading states, fixed video thumbnail distortion, and enhanced design consistency.
Whisper System:
Added rendering for network status changes.
Introduced avatar fallbacks, relative timestamps, and dropdown filtering.
Fixed user auto-selection issues and updated integration to Inbox message API.
Threads (Network):
Introduced Thread Overview Modal for editing thread name, description, and logo.
Added permission-based access, live field validation, and dynamic pinned posts.
Integrated ThreadPostTextarea for improved input, media, and tagging.
Enhanced scrolling, edit/cancel handling, and UI consistency.
Cypress Tests: Added new E2E tests for side menu, onboarding, and session caching; improved test stability and exception handling.
Backend Improvements
Completed reply-to-post flow in Threads module.
Updated Whisper flow — whispers now appear in Inbox chats through global SSE.
Created Perceptron AdvanceInfo v2 flow, maintaining v1 for backward compatibility.
Added AI sync endpoint to maintain user–perceptron memory connection.
Optimized Threads and Network APIs for smoother role-based access and permission validation.
AI Enhancements
Introduced KnowledgeBaseSummary Tool for high-level topic retrieval.
Added structured multi-step workflows for all AI modes (Execution, Simulation, Sensory).
Refined Action Subagent to separate persona-based and knowledge-driven tasks.
Enhanced Knowledge Base Search accuracy with guided reasoning and topic-based responses.
Improved contextual understanding, task relevance, and memory update flows for continuous learning.
Key Fixes
Fixed “Access Denied” issues in Network Threads after backend updates.
Prevented cross-gender asset equips and fixed overflow in Shop chat.
Improved Sidemenu interaction logic — disabled states now show proper popups.
Fixed User Profile null city display; added missing cities for 68 countries.
Moved Onboarding SVGs to correct public path for reliable asset loading.
Centered OTP and Password screens for visual alignment.
Implemented Avatar 3D fallback for CDN reliability.
Fixed auto-increment sequence conflicts in product-related tables.
Filtered perceptron-owner chatrooms with no active messages.
Refactored Knowledge Base Search to work only in Execution Mode for improved accuracy.
QA & Deployment
Frontend & Backend passed all QA tests.
Successful build deployed to the test environment.
