Photorealistic, multi-modal synthetic data, in minutes

Describe a scenario and Stardust renders it: diverse, labeled, photorealistic data with rich time-series metadata. Generate in minutes what a fleet collects in months.

Real data alone will not get you there

// real data today

• ✕Real data is slow and expensive to collect and label
• ✕The long tail is missing: rare, dangerous, and edge-case scenarios never make the dataset
• ✕Sensor coverage is fixed to whatever hardware you flew, so you cannot test what you did not capture

Multi-domain, photorealistic data matching your operational requirements

One scene, rendered across modalities in perfect registration, with ground-truth labels generated automatically for every frame.

• RGB Photorealistic visible spectrum
• IR Infrared and thermal bands
• DEPTH Per-pixel depth and range
• SEGMENTATION + BBOXES Pixel-perfect masks and 2D/3D boxes

One scene, multiple sensors

Render the same world across modalities in perfect registration, so you train and test the whole fusion stack, not just the camera.

• AVAILABLE RGB Photorealistic visible spectrum
• AVAILABLE IR Infrared and thermal bands
• AVAILABLE Depth Per-pixel depth and range
• IN DEVELOPMENT Radar Range-Doppler returns and clutter

Explore a real maritime scene, frame by frame

Every Stardust scene exports to Rerun. Pan and zoom the 3D world, scrub the timeline, and toggle sensor streams in registration. Live, interactive, real Stardust output.

An error occurred during loading:

failed to create wgpu surface: Failed to create surface for any enabled backend: {Gl: InstanceError { message: "canvas.getContext() returned null; webgl2 not available or canvas already in use", source: None }}

Please consider filing a bug report. Thank you!

Realism and diversity with AI post-processing

A learned post-processing pass closes the domain gap and multiplies diversity. One scene, every condition: weather, lighting, and time of day, all photorealistic, so models transfer cleanly to the real world.

A 1,000+ object 3D asset library, and counting

Drag and drop from over a thousand production-ready, physically-accurate 3D assets, each tagged with real-world dimensions. Need something rare? Generate new assets with AI on demand, or request a custom asset from our 3D team.

Container ship 213 × 43 × 52 m

Container ship (Maersk) 400 × 60 × 73 m

Cruise ship 366 × 69 × 78 m

Response boat 14 × 4.7 × 8.1 m

Kayak 5.5 × 2.1 × 0.5 m

Monitoring buoy 1.0 × 1.0 × 1.1 m

Special mark buoy 2.8 × 2.6 × 6.1 m

Cardinal mark buoy 1.3 × 1.3 × 3.0 m

Cessna 208 Caravan 13 × 16 × 4.6 m

Pilatus PC-12 14 × 16 × 4.2 m

Harbor crane 12 × 28 × 29 m

Shipping port 912 × 1979 × 82 m

Built for the hardest domains

• Robotics
• Maritime
• Aerial
• Space
• Off-road

Trained a policy? Evaluate it on Manifold

Stardust trains and tests your perception models. When you are ready to evaluate a full policy, Manifold runs it on any benchmark and ranks it on a shared leaderboard.