See Deployment in Action
Watch how trained policies go from the platform to real robots — deployed, monitored, and ready for production.
Key Benefits
How Deployment fits into your workflow
One-Click Deployment
Deploy trained models to production with a single API call, eliminating complex deployment pipelines.
Flexible Hosting Options
Choose between cloud-based or on-premises deployment depending on your requirements and constraints.
Low-Latency Inference
Optimized runtime for real-time robot control with ultra-low latency for time-sensitive applications.
Auto-Scaling Infrastructure
Automatically scale to handle varying loads, from a single robot to an entire fleet.
Comprehensive Monitoring
Track model performance, resource utilization, and inference metrics in real-time.
Continuous Improvement
Enable ongoing learning and adaptation through feedback loops and data collection from deployed models.
Deploy in Minutes
From trained model to production deployment in just a few steps
Select Your Model
import neuracore as nc # Login to your account nc.login() # List available models policy = nc.policy_local_server(train_run_name="name of your run")
Choose a trained model from your account to deploy.
Connect
# Use the model for inference nc.log_joint_positions(robot.get_joint_positions()) nc.log_rgb(robot.get_rgb()) prediction = policy.predict() action = prediction.outputs[DataType.JOINT_TARGET_POSITIONS] robot.execute_action(action)
Deploy your model and start using it for inference in your robot control loop.