Live production software, reimagined
BackStageOS
One local-first, peer-to-peer operating system for the whole production.
Every department, every stage of a show’s life, on one shared production graph.
The problem
Live production — theatre, events, broadcast — is run on a patchwork of disconnected, proprietary tools.
Every department bought its own island, and nothing shares a single source of truth.
The department silos
- Lighting design and visualisation — WYSIWYG, Capture, grandMA and Eos ecosystems
- Audio — Fairlight, proprietary DAWs and live consoles
- Video and live switching — vision mixers, media servers, vMix and OBS
- Cue playback — QLab and friends
- Management — spreadsheets, WhatsApp, paper call sheets, binder prompt books
- Comms — closed hardware intercom systems
What that costs
- Data is siloed per department. Lighting does not know the set; sound does not know the video cues.
- Costs lock out small companies. A single WYSIWYG subscription is out of reach for many small and independent productions.
- Nothing talks to anything without brittle glue and manual re-entry.
- The paperwork and the show are different worlds, so information never flows from pre-production into the performance and back.
The vision
BackStageOS is one local-first, peer-to-peer operating system for the whole production,
covering every department and every stage of a show’s life.
The Product Family
BackStageOS
The production OS: the whole platform for a show, from pre-production to broadcast.
VenueOS
The venue OS: the software a building runs on — digital signage, box office, EPOS and building systems.
Maestro
A line of hardware consoles for Lighting, Sound Video and Stage Management as well various nodes to the capture and distribution of input and outputs for the mesh.
All three share one core: the graph, the runtime, the mesh and identity.
Guiding principles
Architecture: two planes
The discipline that keeps a system this ambitious finishable is a small number of hard boundaries
— chief among them a separation between what the show should be and what the show does.
Authoring plane — the graph
Execution plane — the runtime
Everything reads from the graph — either live while editing, or as a compiled snapshot while running. That is the difference between a document and a machine.
Surfaces desktop · department consoles · Maestro · mobile
Execution runtime · cue engine · timeline · Show Position · override layer · log
Authoring production graph (CRDT) · core + pluggable domain namespaces
Placement node manifests · intent to capability binding
I/O engines DMX/sACN/Art-Net · audio (AES67/Dante) · video (SDI/NDI) · OSC/MIDI/GPIO · time
Mesh P2P transport · content-addressed storage fabric
Deployment profiles
same code, different scale
| Profile | Shape |
|---|---|
| Solo | One machine runs every engine; outputs via local interfaces and dongles. |
| Console | One appliance runs the show with a purpose-built surface. |
| Venue (hub + nodes) | A hub coordinates; edge nodes provide dedicated outputs. |
| Multi-site | Federated across venues and touring. |
Why it is different
The market does not lack tools — it lacks the integration.
One graph, one runtime, everything designed to work together is the moat, not any single feature.
One shared graph
Lighting, audio, video, stage management and venue are designed to click together. Every department works against the same authored truth.
Local-first,
failover by design
No single point of failure. Network-native output — sACN, AES67, NDI — means any node can drive the pipe, not just the brain.
Adoptable,
then integrated
Use one app on its own; add the rest as progressive enhancement. Nobody has to buy a whole OS to get one tool.
Open everything
Open protocols and formats, and import from proprietary tools wherever possible. Bridge existing open engines before rebuilding them.
Version control
and reuse
Git-like history over the graph, with libraries so cues, palettes, sounds and templates are reused across shows by reference — not copied.
Cue-aware comms
A software intercom matrix where the stage manager’s calls are timecoded against Show Position. The comms are the documentation.
Standards, not silos
BackStageOS speaks open protocols and interchange formats, and never parses proprietary files directly
It uses the source tool’s own exporter or a headless conversion.
| Domain | Open standard |
|---|---|
| Clock / timebase | IEEE 1588 PTP |
| Audio | AES67 / ST 2110-30 (Dante via bridge) |
| Video | SRT / RIST / NDI / ST 2110; WHIP and WHEP for web monitoring |
| Lighting | sACN (ANSI E1.31) / Art-Net |
| Stage and rig geometry | MVR / GDTF |
| Control | OSC / MIDI |
| 3D interchange | USD / glTF / Alembic |
| Video and edit | OpenTimelineIO / EDL / AAF |
Roadmap: four horizons
The roadmap is directional, sequenced by user value and dependency rather than technology fashion.
Each horizon should leave something usable, not just scaffolding. The current phase is ideation.
| Horizon | Focus |
|---|---|
| 1 — The OS core | Prove the spine end-to-end: production graph, the compile and arm contract, the runtime and Show Position, placement, identity, failover, timebase, one surface and one output engine (sACN). |
| 2 — The department apps | Lighting, audio, video, venue design and stage management — individually adoptable, collectively integrated. Bridge-first integration with existing open engines, plus the built-in virtual rig and rendering. |
| 3 — The platform | VenueOS breadth (digital signage, box office, EPOS, building systems), broadcast-out, cross-site federation, and the Maestro console family. |
| 4 — Adjacent markets (gated) | Rendered and virtual shows, virtual sets, games and virtual concerts, newsroom — only once the beachhead and the platform are solid. |
Milestone 0 — the pulse. A single program: one production, one fixture, one cue, one sACN packet, one GO button. Prove the pulse, then grow the body.
The beachhead: an open lighting app
The first product wedge is a lighting design app — an open, local-first replacement for WYSIWYG (Cast) and, by extension, Capture.
Paperwork is universally needed, and the open standards make it tractable: GDTF for fixture personalities, MVR for the rig, sACN / Art-Net for DMX, DXF/DWG for drawings, and IES/LDT for photometry.
The production graph is truth; the CAD document is a projection; the visualiser is a reader.
| Phase | Goal |
|---|---|
| Phase 0 — Spike (done) | Proved reading real venue drawings, parsing GDTF and MVR, emitting sACN and Art-Net, and rendering a beam on wgpu at roughly 590 fps. Settled CADCraft integration: own the app, depend on the foundation, never fork. |
| Phase 1 — Plot and paperwork | Venue to graph, the plot, editable schedules and PDF paperwork. The honest wedge. |
| Phase 2 — Patch and universes | DMX patch, Art-Net and sACN out and in, visualise levels, connect a real console. |
| Phase 3 — The visualiser | 3D venue and set, fixtures with pan, tilt, colour, gobo and beam; haze and camera. Driven by live DMX or the internal runtime. |
| Phase 4 — Show and console integration | Palettes, groups, effects and cue stack on the graph; the visualiser follows Show Position. |
| Phase 5 — Reports depth and rigging | Data sheets, magic sheets, focus charts and followspot sheets; advisory rigging and load. |
| Phase 6 — Collaboration and cross-department | CRDT multi-user editing, share and collaborate, and cue-aware comms. |
A firm boundary: we advise, certified hardware executes
Anything that can hurt someone — motion, loads, crowd, fire — stays in certified hardware. BackStageOS advises, coordinates, visualises and logs, and is never the certified executor.
Status
Current phase: ideation. The vision, architecture, data model, identity, runtime, interop, venue model and roadmap are written. The lighting app’s Phase 0 spikes are complete — real venue drawings read losslessly, GDTF and MVR parsed, sACN and Art-Net emitted, and a volumetric beam rendered on wgpu. No production code has shipped yet.
Work is tracked in the project docs: vision (00), architecture (01), data model (02), identity (03), runtime and failover (04), engines and interop (05), venue management (06), roadmap (07), and the lighting app plan (18).