Skip to main content

Deployment

Congratulations! Your loop is ready to deploy. Let's get it live and making predictions in your applications.

How to Deploy Your Loop​

Deploying your loop is simple and takes just a few steps:

Step 1: Create Your Loop Definition​

After defining your input/output schemas and task type, create your loop definition through the Heimdall dashboard or API. This initializes your adaptive learning model.

Step 2: Initial Training​

Provide initial training data to bootstrap your model. You can start with as few as 10-20 examples, though more diverse examples will improve initial performance.

Step 3: Generate Your API Key​

Navigate to your loop page and click "Generate API Key" to create a secure key for authenticating your API requests.

Save Your API Key

Please save your API key immediately after generation. If you lose it, you'll need to regenerate a new one as API keys cannot be retrieved once generated.

Step 4: Get Your Loop ID​

Find your Loop ID on the loop detail page. You'll need this along with your API key to make predictions.

Step 5: Start Making Predictions​

Your loop is now live and ready to make predictions! Use the API endpoints to:

  • Make predictions with your input data
  • Get feature contributions for explainability
  • Feed results back as training data for continuous improvement

Loop Lifecycle​

Unlike traditional ML models, loops are designed for continuous operation:

  1. Initial Training - Bootstrap with initial examples
  2. Production Predictions - Make predictions in real-time
  3. Continuous Learning - Feed prediction results back as training data
  4. Automatic Improvement - Model adapts without manual retraining

Integration Points​

Your loop can be integrated into:

  • Web Applications - Real-time predictions in user-facing apps
  • Mobile Apps - On-device or server-side predictions
  • Data Pipelines - Batch processing with continuous learning
  • Microservices - API-first architecture integration
  • Event-Driven Systems - Real-time predictions on events

You're All Set!​

Your loop is now live and ready to make predictions. The API Integration tab has everything you need to start using your loop in your applications.

Ready to Integrate?

For detailed API integration examples, see our API Integration Guide. The guide includes sample code in multiple languages and step-by-step instructions for various use cases.

Next Steps​

  • Monitor Performance - Track accuracy and model improvements over time
  • Add Training Data - Continuously improve by feeding back prediction results
  • Scale Usage - Loops handle high-volume prediction requests efficiently
  • Optimize Schemas - Refine input/output schemas based on real-world usage