Serverless Deployment
Deploy AI models as fully managed, auto-scaling serverless instances
Serverless Deployment
Serverless deployments in Unicron are fully managed, auto-scaling instances that require no infrastructure management.
Benefits
Serverless deployments are ideal for:
- Getting started quickly
- Handling variable traffic patterns
- Cost-effective experimentation
- Projects with unpredictable usage patterns
Creating a Serverless Deployment
From the Marketplace
- Navigate to your workspace
- Click on "Deployments" in the navigation menu
- Click "Create Deployment" button in the upper right
- Select the "Marketplace" tab to view available pre-configured services
- Choose a service and follow the configuration steps
Custom Service Deployment
- Navigate to your workspace
- Click on "Deployments" in the navigation menu
- Click "Create Deployment" button
- Select the "Custom Services" tab
- Click "Deploy Custom Service"
- Fill in the required configuration details:
- Service name and description
- Docker image details
- Hardware specifications (CPU, memory, GPU)
- Scaling configuration (min/max replicas)
- Region selection
Pricing
Serverless deployments use a pay-per-use pricing model:
- You pay only for actual usage (per request pricing)
- No minimum fees or upfront costs
- Automatic scaling to zero when not in use
API Usage
Once deployed, each service exposes an API endpoint that can be integrated into your applications:
- Navigate to your deployment:
/workspace/{workspace-slug}/deployments/{deployment-slug}
- Copy the API endpoint URL
- Use the endpoint with appropriate authentication to access your deployed service
Monitoring
You can monitor your serverless deployments through the dashboard, which provides:
- Request volume metrics
- Latency statistics
- Error rates
- Cost tracking
- Scaling events