Open-source drone light-show animation tool for Blender. Design formations and flight paths of any complexity, validate them for safe flight, and export to the data formats used by modern drone-control software.
Status: Alpha — under active development. See project-plan.md for the full roadmap.
Sirius brings professional drone-show choreography into the Blender viewport. It bridges artistic animation and hardware deployment: generate formations from arbitrary 3D objects, animate drones and their LEDs, compute safe transitions between formations, validate against real-world flight constraints, and export for real shows.
Planned (MVP)
- Scalable swarm representation (Point Cloud + instances) for 1,000+ drones
- Takeoff grid / launchpad generation
- Formation generation from any mesh, curve, text, or logo with density and spacing control
- Per-drone and group LED color animation on the timeline
- Automatic formation-to-formation transitions (optimal slot assignment + velocity-profiled motion)
- Collision and feasibility validation (spacing, speed, acceleration, geofence, altitude) with live viewport feedback
- Multi-format export with coordinate-frame and sample-rate conversion
Implemented
- Parametric takeoff grid generation
- Per-drone LED color and emission control
A layered design keeps the hard logic testable and Blender-version-isolated:
- Core — bpy-free data model and algorithms (assignment, collision, feasibility, interpolation, coordinate conversion)
- Blender adapter — thin
bpyintegration (swarm object, Geometry Nodes, handlers, viewport drawing via thegpumodule) - Exporters — format-agnostic trajectory sampler with one writer per target format
- UI — sidebar panels, operators, property groups
Drones are represented as a single Point Cloud (data backbone) with instanced visual proxies and one shared LED material, driven by Geometry Nodes for formation generation and native F-Curves/NLA for animation.
- Blender 4.5 LTS or newer
- Python 3.11+
- Download the latest release
.zip. - In Blender, open Edit → Preferences → Get Extensions → Install from disk and select the
.zip. - Enable Sirius.
| Format | Type | Target stack |
|---|---|---|
| CSV | Flight-ready | Generic |
| Vimdrones raw | Flight-ready | Vimdrones GCS |
| UgCS PATH / PATH3 | Flight-ready | SPH Engineering Drone Show Software |
| VVIZ | Visualization/interchange | Verge Aero, Finale3D, Depence, FWSim |
| Depence | Visualization | Syncronorm Depence |
The plan is phased around a minimum end-to-end MVP (create → animate → transition → validate → export), followed by accelerators and polish.
- Phase 0 — Stabilize, restructure, scalable foundation
- Phase 1 — Launchpad and formation generation
- Phase 2 — Animation and LED
- Phase 3 — Transitions and assignment
- Phase 4 — Collision and feasibility validation
- Phase 5 — Export engine
- Phase 6 — Accelerators (geo-referencing, pyro, music sync, flocking, import round-trip) — post-MVP
- Phase 7 — Polish and release
Full details, including verified Blender API decisions and the test strategy, are in project-plan.md.
Contributions are welcome. Pick an open issue labeled phase-N or good first issue, and read project-plan.md for architecture and conventions before starting. Core algorithm and exporter modules are pure Python and unit-testable without Blender.