Changelog

What's New

Every update, improvement, and architectural milestone.

February 2026

v1.1 — Auth, Plans & Marketing Site

JWT authentication with bcrypt password hashing. Three-tier plan system (Free/Pro/Enterprise) with server-side enforcement. Marketing website with 18 pages. Intent detection for Direct vs Sprint mode routing. Self-model service with execution outcome tracking, tool reliability monitoring, user correction learning, and dynamic system prompt addendum. 8-layer memory backed by SQLite with per-layer TTL. Community Learning opt-in framework. Download page for desktop app.

February 2026

v1.0 — Public Beta Launch

Six stations live: Chat, Code Lab, TBWO Command, Research Hub, Image Studio, Voice Room. Server backbone with SQLite persistence (40+ REST endpoints). SSE streaming proxy for Anthropic and OpenAI (API keys server-side only). TBWO execution engine with parallel pod scheduling, checkpoint gating, and receipt generation. Domain-specific pod prompts for 8 project types. Capability detection for web vs desktop feature gating.

January 2026

v0.9 — Architecture Complete

64,000+ lines of TypeScript across 214 files. Domain-specific prompts with getDomainPodPrompt() combining base role, domain instructions, objective-derived context, and quality tier. Message bus for inter-pod communication with pub/sub, correlation IDs, and priority levels. Hardware dashboard with real-time CPU/GPU/RAM monitoring. Home Dashboard as hub-and-spoke mission control with live station previews and keyboard shortcuts.

December 2025

v0.8 — TBWO Execution Engine

TBWOExecutor (3,654 lines): parallel pod scheduling, checkpoint gating with configurable intervals, artifact bus for inter-pod file passing, time-scaling plan generator with 4 complexity tiers, and receipt generation. Pod system with 7 specialized roles (Orchestrator, Design, Frontend, Backend, QA, Copy, Research). Resource pool with concurrent pod limits.

November 2025

v0.7 — Memory & Voice

MemoryService (2,908 lines): 8-layer cognitive memory with per-layer decay rules, salience scoring, and consolidation. Whisper speech-to-text and ElevenLabs text-to-speech integration. Emotion-aware voice synthesis with adaptive pitch, speed, and tone modulation. ClaudeClient with extended thinking support, streaming, retry logic, and cost tracking.

October 2025

v0.5 — Core Platform

React + TypeScript + Vite foundation. Zustand state management with persistent stores. Chat with multi-model streaming (Claude + OpenAI). Sidebar with conversation management. Settings modal. Brave Search integration. File upload and preview. Code block rendering with syntax highlighting. Dark cinematic UI with the ALIN design system.