Modal has raised an $87M Series B led by Lux Capital. Read more

Modal Sandboxes

Secure Sandboxes at massive scale

Run AI-generated code in dynamically defined sandboxes. Instantly scale to 50,000+ concurrent sessions.

“Modal was the only infrastructure provider that enabled us to reliably run tens of thousands of app creation sessions in an instant. We’re excited to build with them for the long term.”

Anton Osika, Founder and CEO

“Everyone here loves Modal because it helps us move so much faster. We rely on it to handle massive spikes in volume for evals, RL environments, and MCP servers. Whenever a team asks about compute, we tell them to use Modal.”

Aakash Sabharwal, VP of Engineering

DYNAMICALLY DEFINED

CODE-FIRST INTERFACE

SUB-SECOND STARTUP

MASSIVE SCALE

Modal Sandboxes

You create Sandboxes, we handle
resource orchestration

Dynamically defined

Create a sandbox with one line of code. You can dynamically define the sandbox environment at runtime.

sb = modal.Sandbox.create(

    app=app,

    image=modal.Image.debian_slim()

        .uv_pip_install("numpy"),

Code-first interface

Send commands to the sandbox from your application code.

p = sb.exec(

    "python",

    "-c",

    "import numpy; print(numpy.__version__)"

print(p.stdout.read())

sb.exec(

    "rm",

    "-rf",

    "--no-preserve-root",

"/"

Sub-second startup

Modal’s container stack spins up sandboxes in less than a second.

Massive scale

Instantly autoscale to 50,000+ sandboxes during peak demand.

Robust primitives to integrate with your application

Snapshotting

Save and restore Sandbox filesystem and memory state.

Tunnels

Directly connect to a Sandbox with a URL.

Networking controls

Set granular restrictions on outbound network access for security.

Filesystem APIs

Easily sync data into and out of Sandboxes.

Fine-grained observability

Debug fast with our native observability interface.

Easily drill down into metrics, logs, and status for individual Sandboxes.

Debug fast with our native observability interface.

Easily drill down into metrics, logs, and status for individual Sandboxes.

Modal Sandbox Pricing

Only pay for what you use, by the CPU cycle. For GPU Sandboxes, refer to our standard GPU prices.

Compute costs

CPU

Physical core
(2 vCPU equivalent)

$0.00003942 / core / sec

*minimum of 0.125 cores per container

Memory

$0.00000672 / GiB / sec

Built with Modal

All examples

Sandbox a LangGraph agent's code

Run an LLM coding agent that runs its own language models

Build a stateful, sandboxed code interpreter

Execute untrusted Python code in a sandboxed environment

Safely run untrusted Node.js, Ruby, PHP, and more

Execute arbitrary code in your language of choice inside a Sandbox

Spawn RL environments for a coding model

Use GRPO with TRL to spawn reinforcement learning environments for training coding agents

Your end-to-end ML lifecyle in one place

Seamlessly integrate data pre-processing, training, and serving.

Ship your first app in minutes.

Get Started

$30 / month free compute