SPACE

What is SPACE?

SPACE, short for Sandboxed Platform for Agentic Code Execution, is an open-source sandbox platform published by Perplexity. Your application calls its HTTP API to create a Linux sandbox from a template, start commands, read their output and exit results, and manage the sandbox lifecycle. Each sandbox is a Firecracker microVM with its own writable filesystem, and several commands can run in the same one, which suits AI agents, per-user workspaces and code-running jobs. Lifecycle changes are asynchronous, so the API returns an operation ID to poll, while command output travels over a separate connection protected by a session token. Sandboxes can be kept, paused or suspended and woken later, and warm pools and base snapshots reduce cold-boot time. An E2B compatibility wrapper lets the tested E2B Python SDK talk to it. SPACE is written in Rust, licensed under Apache 2.0, and you deploy and operate it yourself, starting with a local Kind cluster; no hosted price is published.

Use cases

• Run AI agent code in isolation
• Give each user a sandboxed workspace
• Execute untrusted code safely
• Stream command output to an app
• Pause and resume sandboxes
• Self-host a code execution service

Who it's for

• AI agent developers
• Platform engineers
• Infrastructure teams
• Backend developers

Why use SPACE

• Each sandbox is a separate microVM, so code from an agent or a user runs apart from your application.
• The HTTP API means you need no Rust library to use it, and the first-sandbox guide uses plain Python.
• Pause, suspend and warm pools help control start-up time and idle cost.
• The E2B compatibility wrapper may ease moving from an existing E2B-based setup.
• Open-source code under Apache 2.0 lets you inspect and modify the platform.

SPACE features

• Firecracker microVM sandboxes
• HTTP API for sandbox and command control
• Templates as starting environments
• Multiple commands per sandbox
• Asynchronous operations with operation IDs
• Separate data plane for command output
• Keep, pause and suspend lifecycle options
• Warm pools and base snapshots
• E2B compatibility wrapper for the E2B Python SDK
• Operator CLI and observability options

SPACE integrations

• Kubernetes
• E2B Python SDK
• Firecracker
• AWS EKS

How to get started with SPACE

1. Deploy SPACE by following the installation guide, starting with a local Kind cluster.
2. Authenticate and create a sandbox from a template through the HTTP API.
3. Poll the returned operation ID until the sandbox is ready.
4. Create a command session, attach to start the process, and read output and the exit result.
5. Keep, pause or delete the sandbox when the job finishes.

SPACE limitations

• You deploy and operate the platform yourself
• Setup documentation covers Kubernetes, starting with a local Kind cluster
• Fields marked incubating are outside the supported contract
• Lifecycle calls are asynchronous and need polling

SPACE pricing

SPACE is open source, so you can use and self-host it for free.

Prices can change and may vary by country or exclude tax. Check the official pricing page.

SPACE alternatives

The main SPACE alternatives listed on Debuly are E2B, Modal, Runloop, CubeSandbox, AgentScope and Busabase. The table compares pricing and platforms; order is A–Z.

1. 

   AgentScope

   Python agent framework with an SDK, web UI and multi-tenant backend for building, sandboxing and deploying AI agents.

   Developer platformsAutomation and agentsCoding
   Open sourceVisit AgentScope (official site)
2. 

   Busabase

   MIT-licensed database and workspace where AI agents store data, docs, skills and apps, available on desktop, self-hosted or in Busabase Cloud.

   Developer platformsAutomation and agentsProductivity
   Open sourceVisit Busabase (official site)
3. 

   CubeSandbox

   Self-hosted MicroVM sandbox service for AI agents, with fast starts, hardware isolation and E2B SDK compatibility.

   Developer platformsAutomation and agentsCoding
   Open sourceVisit CubeSandbox (official site)
4. 

   E2B

   Cloud sandboxes where AI agents run code safely, with Python and JavaScript SDKs and per-second usage billing.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit E2B (official site)
5. 

   Modal

   Serverless cloud platform for running AI inference, training, batch jobs and agent sandboxes, billed per second of compute.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit Modal (official site)
6. 

   Runloop

   Sandboxed devboxes, blueprints and snapshots for building and running AI software engineering agents.

   Developer platformsCodingAutomation and agents
   FreemiumVisit Runloop (official site)

SPACE FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from avatars.githubusercontent.com · fetched 24 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.