Modal
Python-native serverless compute — GPUs, batch jobs and agent sandboxes, billed per second
About
Modal is a Python-native serverless compute platform: you decorate a Python function, declare its container image and GPU in code, and modal deploy turns it into an autoscaling cloud endpoint — no Docker, YAML or Kubernetes. It covers GPU/CPU batch jobs, LLM inference with memory-snapshot fast cold starts, scheduled jobs, notebooks, and a Sandboxes API for executing agent-generated code, with per-second billing pooled across 13+ cloud providers.
Modal Labs is heavily independently covered: a $355M Series C (May 2026) at a $4.65B valuation, ~$300M annualized revenue, and SOC 2 Type II since January 2025. Only the client SDKs are open source (Apache-2.0); the execution engine itself is closed, so Modal cannot be self-hosted.
What it can do
- • Python functions as autoscaling endpoints via decorators — image, GPU and dependencies declared in code
- • GPU workloads (A100/H100/B200-class cards) plus CPU batch jobs, scheduled jobs and GPU notebooks
- • Memory snapshots for fast cold starts (sub-second cold-start figures are provider-claimed)
- • Sandboxes API: thousands of isolated environments for executing agent-generated code
- • Volumes, secrets, and TypeScript/Go SDKs for calling Modal functions cross-language
- • Per-second billing with a free starter tier (provider-claimed); SOC 2 Type II, HIPAA via Enterprise BAA
Setup
- Create an account at https://modal.com.
- Install the package:
pip install modal. - Authenticate:
modal setup(browser-based; if it fails, trypython -m modal setup). - Write a function in a Python file and run it:
modal run path/to/file.py. - Deploy it as a persistent app/endpoint:
modal deploy path/to/file.py.
Prerequisites
Modal account + `modal setup` authentication; Python 3.8+. A valid payment method is required for GPU workloads (per Modal docs).
Use cases
Works with
Trust score — 7/10
Scored 7/10 (verified tier, metrics-v1). Evidence: 523 GitHub stars (log-scaled vs catalog peers); blended with ~47,839,937 monthly package downloads; last push 0d ago; 0 releases/12mo; 32 open issues. Manual review: core engine closed (no self-host); GPU workloads need payment method. License: Apache-2.0 SDK; core engine closed-source.
★ 523 GitHub stars
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 6.8/10
- Maintenance · recency, releases, issue load
- 6.1/10
- Openness · license + docs
- 8.6/10
- Manual review · sanity read + red-flag check
- 9.0/10
- · 523 GitHub stars (log-scaled vs catalog peers)
- · blended with ~47,839,937 monthly package downloads
- · last push 0d ago; 0 releases/12mo; 32 open issues
- · manual review: core engine closed (no self-host); GPU workloads need payment method
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.