The Voice-Driven Command & Creation Substrate.
The engineering laboratory and software architecture of Justin Andrew Wood. Built to emancipate human focus through voice-driven command pipelines, zero-token daemon fleets, and automated procedural video and audio synthesis.
From Audio Terminal to Autonomous Video & Voice Runtime
Started as an Electron-based audio generation terminal application, featuring real-time Web Audio 2.0 DSP synthesis, keyboard-first soundboard, and ambient drone engines.
Axis Mundi Bridge
Zero-token Google Keep synchronization bridge, SSE live event streams, Telegram daily dispatches, and local Model Context Protocol (MCP) tool registry.
Foundations Orchestrator
Industrial multi-pipeline Go orchestration chassis. Translates triaged voice intent into code compilation (CODE_DEV), Google Workspace actions (STRATEGIC_PLAN), and procedural storyboards (CREATIVE_STORY).
AMRA Video Studio
Automated 3-tier video generation fleet: Mobile Montage (Mode 1), Procedural Sacred Geometry (Mode 2), and Bicameral Spoken-Voice & Relic Synthesis with kinetic ASS subtitles (Mode 3).
Continuous Voice Perception & Sovereign Observer
The continuous cognitive loop. Google Recorder speaker diarization (:8091), fast-whisper transcription (:8095), local Ollama Moondream VLM image classification (:8086), and the autonomous 5-second Sovereign Observer executive loop (:8092) writing to SQLite WAL.
Automated Cinema-Grade Video Generation
AMRA transforms raw spoken-voice recordings and diarized transcripts into vertical 9:16 Shorts or widescreen 16:9 master productions. Featuring multi-axis Ken Burns rendering across 5,800+ relics, in-memory 48kHz audio beds (-14 LUFS), and kinetic ASS subtitles with platform safe zones.
Pixel 7 clip stitching, xfade transitions, and telemetric lower-thirds.
Procedural sacred geometry, Saturnian toroids, and pure Go 48kHz WAV audio.
Bicameral spoken-voice synthesis, kinetic subtitles, and Ken Burns animation.
Headless Modes 1/2/3, pure Go 48kHz WAV audio synthesis & kinetic subtitle burn-in.
Multi-pipeline Go orchestrator: CODE_DEV compiler, workspace admin & storyboards.
High-speed voice ingestion daemon, MCP tool protocol, and daily telegram digest.
Ephemeris modeling, photo ingestion pipeline & 85,000-file storehouse indexer.
Zero-cloud Ollama Moondream VLM image intake and intelligent ontology categorization.
Continuous 5s executive loop, 12-sensor Sensorium & 102,000+ logged SQLite cycles.