Your account is pending approval. You'll receive an email at when your account is activated. This usually takes less than 24 hours.

Simulacrum Labs — Render at Scale

Connecting...
Docs
Upload Your UE5 Project
📦
Drop .zip or .7z here or click to browse
Packaged UE5 project (.zip, .7z) — any size (files >2 GB upload direct to Azure)
Uploading...
Configure Render Job
Inject annotation capture into client project
Labels JSON + COCO manifest auto-included with every job
Environment Overrides (Optional) ▶
Vehicle type for AirSim projects. Choose "None" for non-AirSim projects.
API = Python script drives the vehicle. RC = manual joystick/keyboard control.
Path relative to project root
Optional extra arguments passed to the script
Path relative to project root. Defaults to main.py if empty.
Optional extra arguments passed to the script
No UE5 Required
Runs entirely inside a Docker container with GPU passthrough. Ideal for IsaacSim, Gazebo, PyBullet, or any Python/GPU workload.

Each parameter becomes a CLI argument passed to your training script. The sweep will run every combination of values across all parameters.

Argument name Values (comma-separated)
Total training runs: 0
Show example
Example: Sweep learning rate and batch size (2 × 3 = 6 training runs)
--learning_rate → 0.0001, 0.001, 0.01
--batch_size → 32, 64
This generates 6 jobs, each with a different combination:
Run 1: --learning_rate 0.0001 --batch_size 32
Run 2: --learning_rate 0.0001 --batch_size 64
Run 3: --learning_rate 0.001 --batch_size 32
... and so on
Active Sweeps
Jobs
Job ID Project Status Frames Progress
No jobs yet. Upload a project and submit a render job.
Activity Log
Credits
$-.--
Loading...
Transaction History
DateTypeAmount
API Keys
Preview
No preview
-
Download
Downloading...