MAESTRA Monitor

// Fleet Slots

0 entities

// Audio Analysis

// Frequency Bands

Sub

Bass

Mid

High

// Stems from TD / OSC

Drums

Bass

Melody

Vocals

// Live Readings

BPM

RMS

time	sub	bass	mid	high	rms

Select a slot

Live

No slot selected

// Transcription

waiting for speech...

Nouns 0 extracted

none yet

Base Prompt

P weight 1.0

// Event Log

// WebSocket Log

// How You'd Use This

Multi-Screen VJ Set

Each screen runs its own TD + Scope instance doing real-time AI generation. Maestra syncs scene changes, prompt updates, and color palettes across every machine. One state push switches the whole venue.

Gallery Installation

Register each station as an entity. Sync brightness based on time of day, coordinate color transitions, advertise NDI streams so curators can monitor every feed from a single dashboard.

Festival Projection Mapping

Multiple building facades with different TD machines. Maestra manages scene changes across the fleet — one API call switches every projector to the next act. Heartbeats alert you if any machine drops.

    # In TouchDesigner  after dropping the Maestra TOX into your project:

    maestra = op('maestra').ext.MaestraExt

    # Register this machine with the fleet

    maestra.Connect()  # uses Entity ID + Server URL from custom pars

    # Push state  every connected client sees the update

    maestra.UpdateState({'brightness': 75, 'scene': 2, 'color': [0.8, 0.2, 1.0]})

    # Advertise your NDI output so other machines can discover it

    maestra.AdvertiseStream(name='Stage Visuals', stream_type='ndi', protocol='ndi')

    # From another machine  discover what's available

    streams = maestra.ListStreams(stream_type='ndi')

// Cloud Nodes � Scope / Daydream API

If you're using Scope or the Daydream API for your project, use this tab to test and select your connection before show day. Decentralized GPU routing can vary � this lets you audition nodes day-of and bring the best ones into the main dashboard, so you're not discovering problems during the opening.

// GPU Node Selector

Preview each node � select the best one to lock in as your active source

With decentralized GPU, once you commit to a node you're locked to that instance. Preview all available nodes here before committing � check stream quality, latency, and FPS. Hit Set Active to route that node's output to your main Krista1 slot.

How Scope Fits In

Scope runs on a single machine — it takes a video source and transforms it in real-time using AI diffusion models. Maestra runs across machines — it coordinates state and streams between devices. They solve different problems, and together they let you run coordinated AI-generated visuals across an entire venue.

                       MAESTRA SERVER
        state sync  stream registry  heartbeat monitor
             REST API    NATS    MQTT    WebSocket

 MACHINE A              MACHINE B              MACHINE C

  Scope                  Scope                  TouchDesigner
  (AI diffusion)         (AI diffusion)         (generative)
      NDI out                    NDI out                    NDI out

  TouchDesigner         TouchDesigner         projector out
  + Maestra TOX         + Maestra TOX         + Maestra TOX
  projector out          projector out

Scope Generates Locally

Each machine runs Scope with its own GPU. Camera or media in, AI-transformed video out. Scope outputs via NDI — TouchDesigner picks it up with an NDI In TOP. Scope doesn't know about other machines. It doesn't need to.

Maestra Coordinates the Fleet

The TOX in each TD instance registers with the Maestra server. When you push a state change — scene: 3, color: [0.8, 0.2, 1.0] — every connected machine receives it. TD reads the new state and updates Scope's prompt, switches LoRAs, or adjusts post-processing.

Streams Stay Discoverable

Each machine advertises its NDI output through Maestra's stream registry. A monitor station can discover all streams on the network — no manual IP hunting. Heartbeats tell you instantly if a machine drops.

    # Example: Maestra state change triggers Scope prompt update via OSC

    def onValueChange(par, prev):

        if par.name == 'Scene':

            scene = par.eval()

            prompts = {0: 'ethereal crystal cave, bioluminescent', 1: 'baroque gold cathedral, dramatic lighting', 2: 'neon cyberpunk rain, reflective surfaces'}

            op('scope_osc_out').sendOSC('/scope/prompt', prompts.get(scene, ''))

Custom Parameters

Connection

Entity ID
Server URL
API Key
Auto Connect
Connect / Disconnect
Status (read-only)
Heartbeat Interval

State

Brightness (0–100)
Opacity (0–1)
Color (RGB)
Speed (0–10)
Scene (int)
Custom State JSON
Push / Pull State

Streams

Stream Name
Stream Type
NDI / Syphon / Spout / RTMP / SRT / WebRTC
Protocol
Address / Port
Advertise / List Streams
Active Stream ID

Gateways

OSC In / Out Port
OSC Prefix
WebSocket Enable / Port
MQTT Enable / Broker / Port

Internal Operators

MaestraExt

Python extension with Connect(), Disconnect(), UpdateState(), PushCurrentState(), PullState(), AdvertiseStream(), ListStreams(), StreamHeartbeat(). Auto-connects on start if enabled. State params auto-push on value change with debounce.

Infrastructure

heartbeat_timer — Timer CHOP, auto-heartbeat every N seconds
par_exec — Parameter Execute DAT
log — Table DAT, rolling activity log
streams_table — Table DAT, discovered streams
info — Info CHOP, component monitoring

Resources

Download TOX GitHub README Setup Guide

    # Quick start:

    maestra = op('maestra').ext.MaestraExt

    maestra.UpdateState({'brightness': 50, 'scene': 2})

    maestra.AdvertiseStream(name='My NDI', stream_type='ndi')

    streams = maestra.ListStreams()