Does Modal support persistent storage?

Yes, Modal Volumes provide persistent storage across function invocations, making it easy to store model weights, cache datasets, or share data between functions.

How does Modal handle autoscaling?

Modal automatically scales functions from zero to thousands of containers based on demand, and you can configure specific limits using max_containers and min_containers parameters.

What is the Modal skill for Claude Code?

The Modal skill provides domain-specific guidance for using the Modal platform to run serverless Python code, specifically optimized for machine learning and heavy batch workloads.

Can I run functions on GPUs with Modal?

Yes, Modal supports a wide range of GPUs including T4, L4, A100, H100, and the flagship B200 for high-performance training and inference.

How do I authenticate with Modal using this skill?

You can use the CLI command 'modal token new' to authenticate, which opens a browser to create a token stored in your local configuration.

Modal Serverless Python

Name: Modal Serverless Python
Author: x-cmd

byx-cmd

•

클라우드 인프라

Runs Python code in the cloud with serverless containers, high-performance GPUs, and automatic scaling for AI/ML workloads.

소개

Modal is a powerful serverless platform that allows developers to execute Python functions in the cloud without managing infrastructure. This skill enables Claude to help you define container images, request specific GPU hardware like H100s or A100s, and scale compute-intensive tasks from zero to thousands of containers. It is ideal for deploying machine learning models, running massive batch processing jobs, or building scalable serverless APIs that require specialized hardware acceleration.

주요 기능

Automatic scaling from zero to thousands of concurrent containers
On-demand GPU acceleration including H100, A100, and L40S types
Persistent storage via Volumes for model weights and datasets
8 GitHub stars
Declarative container image building with custom dependencies
Serverless web endpoints and scheduled cron job management

사용 사례

Training machine learning models on high-performance GPU clusters
Processing massive datasets in parallel using distributed compute
Deploying and scaling LLM inference or image generation models

소개

주요 기능

Automatic scaling from zero to thousands of concurrent containers
On-demand GPU acceleration including H100, A100, and L40S types
Persistent storage via Volumes for model weights and datasets
8 GitHub stars
Declarative container image building with custom dependencies
Serverless web endpoints and scheduled cron job management

사용 사례

Training machine learning models on high-performance GPU clusters
Processing massive datasets in parallel using distributed compute
Deploying and scaling LLM inference or image generation models