# Dagster + Modal

> Run serverless Python functions at scale with Modal.

Source: https://dagster.io/integrations/dagster-modal
Category: Compute

## Dagster & Modal

## About this integration

Modal is a serverless cloud platform that makes it easy to run Python code at scale without managing infrastructure. The Dagster-Modal integration enables you to execute compute-intensive workloads, ML training jobs, and data processing tasks on Modal's auto-scaling platform directly from your Dagster pipelines.

With this integration, you can:

- **Run serverless functions** that automatically scale from zero to thousands of containers
- **Access GPU resources** for machine learning training and inference workloads
- **Execute long-running jobs** without worrying about infrastructure management
- **Optimize costs** with pay-per-use pricing and automatic scaling
- **Deploy instantly** with Modal's container orchestration
- **Monitor execution** through Dagster's observability tools

Modal is particularly powerful for ML workflows, data processing at scale, and any compute-intensive tasks that benefit from elastic infrastructure.

## Installation

```bash
pip install dagster-modal modal
```

## Key Features

### Serverless Execution

Modal automatically provisions and scales compute resources based on demand, eliminating the need to manage servers or containers.

### GPU Support

Access powerful GPU instances (A10G, A100, H100) for machine learning training and inference without long-term commitments.

### Instant Deployment

Deploy functions and containers instantly with Modal's fast cold-start times and efficient resource allocation.

### Cost Optimization

Pay only for actual compute time with automatic scaling to zero when idle, making it cost-effective for variable workloads.

### Container Orchestration

Run complex applications with full control over the runtime environment, dependencies, and resource requirements.