From the archive
Modal opens general access to its serverless Python platform
A serverless platform runs Python functions, GPU inference, and batch jobs in the cloud with usage-based billing.
Modal opened general availability of its cloud function platform, removing the waitlist for account registration. The service runs Python workloads in hosted containers, including GPU inference and large batch jobs.
Infrastructure described beside the code
Modal lets a Python application define the container image, hardware, and persistent storage it needs. Functions can become web endpoints or scheduled jobs, while inference endpoints can scale with requests.
The launch's examples include transcription, image generation, and other workloads that need more compute than a local development machine.
Pay for running work
Modal hosts the execution environment and charges per second of use. That gives intermittent workloads a different cost structure from keeping a dedicated server available throughout the day. The platform also supports parallel batch work across many containers.
The Modal profile covers its serverless model, graphics hardware options, and deployment workflow.