Ecosystem roadmap

[!NOTE] This page describes the ecosystem-wide trajectory. For the version-by-version Pit roadmap with exit criteria per release, see Pit roadmap.

Where we are

Layer	Status
Stardust Pit	🟢 In active development — v0.5.0 (multi-plugin chain hosting) shipped; v0.6.0 engine completeness is next
Stardust Sheets	⚪ Design intent only — starts post-Pit v1.0
stardust-core shared library	🟢 v0.5.0 — patch graph, show document, CLAP host, 3-band EQ + StereoChannel, audio/MIDI primitives
stardust-workspace meta-repo	✅ Set up with ADRs, justfile, scripts, CLAUDE.md
stardustmt.github.io (this site)	✅ Live
Speculative future apps	⚪ Named, not built

Trajectory

Now ──── Pit v0.x ──── Pit v1.0 ──── Sheets v0.x ──── Sheets v1.0 ──── (eval ecosystem expansion)

Two products earning their audience is the gate. The rest of the ecosystem is sized — and only built — based on whether Pit and Sheets land.

Phase 1 — Pit reaches public release (~now → v1.0)

Build Pit until it’s reliable enough to use as a daily driver in a professional pit. The version-by-version plan with exit criteria lives on the Pit roadmap; the headline arc is:

• v0.1.0 – v0.5.0 ✅ shipped — Tauri scaffold, patch editor, engine, show document, on-screen MIDI, always-on engine, multi-plugin chain hosting
• v0.6.0 (next) — Engine completeness: hardware MIDI bindings, plugin GUI hosting, scan caching, PR CI
• v0.7.0 — Plugin sandboxing (out-of-process), watchdog, crash recovery, soak tests
• v0.8.0 — Transport + MD essentials: per-bar tempo, click bus, MIDI clock send, tap tempo
• v0.9.0 — Three-mode shell, splash, New Show wizard, settings window
• v0.10.0 — Library data model unification, drawing primitives, Pit Mixer
• v0.11.0 — Perform mode + widget catalog + conductor cam
• v0.12.0 — Click track editor + LUFS balance tool
• v0.13.0 — Backing tracks + .stardustshow bundle format + MIDI recording
• v0.14.0 — Native SFZ player + bundled GM piano
• v0.15.0 — Polish, signed installers, extension API, accessibility audit, tech-debt sweep
• v1.0.0 — Open beta → public launch with auto-update

Versions are framed as a release schedule under semver, not as “phases” — every version is a real release with exit criteria, not a planning convenience.

Phase 2 — Sheets starts (post-Pit v1.0)

Sheets is a substantial second product, deliberately not started until Pit is shipping and stable. The general arc:

• Define the semantic overlay format and migration story
• Build the iPad-first reader experience
• Add the Windows + macOS editors
• Production-aware library + distribution
• v1.0 when Sheets can replace forScore for a working musical theatre pit player

See Sheets design intent for the day-one philosophy.

Phase 3+ — Ecosystem expansion (speculative)

If Pit and Sheets find their audience, the platform expands. The candidate apps are named so the direction is visible — none are committed.

• Stardust Rehearse — production participant companion: schedules, distributed materials, rehearsal notes, cast coordination. Mobile-first.
• Stardust Produce — creative production collaboration: blocking, staging, design collaboration, scene planning. Consumes content from a future Write / Compose / Arrange family.
• Stardust Stage — stage management runtime: backstage operations, cue tracking, show running, reports, comm. Distinct from Pit (musician runtime) and Produce (creative pre-production).
• Stardust Lighting — lighting runtime and control: DMX, Art-Net, sACN, OSC, MSC.
• Stardust Galaxy — optional cloud layer: sync, marketplace, collaboration, plugin ecosystem. Local-first remains absolute; Galaxy is never required.

The order in which these get built (if any do) depends entirely on what Pit and Sheets reveal about user needs. Listing them isn’t a roadmap commitment; it’s a way of being explicit about where the platform could go so the foundational decisions in Pit and core don’t accidentally close those doors.

What’s explicitly off the table

• Replacing notation software (MuseScore, Sibelius, Dorico). Sheets layers metadata on PDFs.
• Full OMR (optical music recognition). The semantic overlay approach sidesteps it.
• Enterprise theatre ERP. If only a 50-person production company would use it, it doesn’t ship.
• Mandatory cloud / subscription / DRM. Apps work fully offline forever.
• AI features without theatre-specific value. AI for AI’s sake doesn’t ship.
• A separate stardust-ui library before Sheets exists. Premature shared-UI extraction.
• Plugin ABI stabilization before there are real plugin authors.

How decisions get made

Significant architecture changes require an ADR. Schemas are versioned (ADR-0003). Cross-app protocols carry version discipline. PRs land small. The principles outlive any individual feature.