E2B

What is E2B?

E2B provides isolated cloud sandboxes that AI agents and apps use to run code. Through the Python and JavaScript SDKs, a CLI and an API, developers create a sandbox, execute code in it, and pause or kill it when the job ends. Sandboxes start from templates that you can customize with your own CPU and memory settings, and a Code Interpreter service runs inside them. Developers building coding agents, data analysis tools and agent back ends use it to keep generated code away from their own servers. Billing is usage-based, charged per second while a sandbox runs, with paused and killed sandboxes costing nothing. The Hobby plan has a $0 base price and a one-time $100 of free credits, Pro is $150 per month with higher limits, and Enterprise is custom. Sandboxes are CPU-only, with no GPU option, and an EU cluster is available. E2B reports a SOC 2 Type II report through its Trust Center.

Use cases

• Run AI-generated code in isolation
• Execute data analysis for an agent
• Give coding agents a safe workspace
• Build custom sandbox templates
• Pause and resume long agent sessions
• Track sandbox lifecycle with webhooks

Who it's for

• AI engineers
• Agent developers
• Startups building AI apps
• Enterprise platform teams

Why use E2B

• Generated code runs in a separate sandbox, so a faulty or hostile script does not touch your own infrastructure.
• Per-second billing for running sandboxes only means idle or paused sessions stop costing money.
• The Hobby plan includes $100 of one-time credits, which lets a team test an agent before paying.
• Pause and resume keep state for later, which suits agents that work across long sessions.
• Custom templates let you choose vCPU and memory per workload instead of a fixed size.

E2B features

• Isolated sandboxes created through Python and JavaScript SDKs
• Custom templates with set vCPU and memory
• Pause and resume with paused sandboxes kept indefinitely
• Code Interpreter service inside the sandbox
• Lifecycle event webhooks for created, paused and killed sandboxes
• CLI and console for tracking active sandboxes and usage
• Public templates shared across projects
• Spending limits and a usage tab in the dashboard
• Secrets stored per project
• Shared EU cluster for verified workspaces

E2B integrations

• Docker

How to get started with E2B

1. Install the E2B Python or JavaScript SDK and create an API key in the console.
2. Create a sandbox from the default or a custom template, which starts with 2 vCPU and 512 MiB of RAM.
3. Send code or commands to the sandbox and read back the output from your agent.
4. Pause the sandbox to stop billing while keeping its state, or kill it to release resources.
5. Watch usage in the console and set a spending limit on the budget page.

E2B limitations

• Sandboxes are CPU-only, with no GPU option
• Hobby sandboxes run for up to 1 hour continuously and allow 20 concurrent sandboxes
• Pro does not add free credits, only higher limits
• Volumes are in beta, with file locking that can hang and no snapshots or read-only mounts yet

E2B pricing

Yes, E2B has a free plan. Paid plans start at $150 per month, according to the official pricing page on 24 Oct 2026.

E2B plans

• Hobby

  $0/month

  plus usage; $100 one-time free credits

  • 8 vCPUs and 8 GiB memory max
  • 10 GiB disk
  • 1 hour max continuous runtime
  • 20 concurrent sandboxes
  • 10 secrets per project

• Pro

  $150/month

  plus usage

  • 8+ vCPUs and 8+ GiB memory
  • 20+ GiB disk
  • 24 hours max continuous runtime
  • 100 - 1,100 concurrent sandboxes
  • 100 secrets per project

• Enterprise

  Custom

  • Custom limits and rate limits
  • 1,100+ concurrent sandboxes
  • Volume commitments for lower prices
  • Contact sales

Plans as listed on the official pricing page on 24 Oct 2026. Prices can change and may vary by country or exclude tax. Check the official pricing page.

E2B alternatives

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

1. 

   Letta

   Platform, formerly MemGPT, for building stateful AI agents with persistent memory, from a terminal app, SDK or API.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit Letta (official site)
2. 

   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)
3. 

   numbat

   Open-source endpoint tool that watches AI agent activity, flags risky behavior locally and can optionally block actions.

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

   ReMe

   Open-source memory layer for AI agents that stores knowledge as local Markdown files with hybrid search and a linked graph.

   Developer platformsAutomation and agentsCoding
   Open sourceVisit ReMe (official site)
5. 

   Runloop

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

   Developer platformsCodingAutomation and agents
   FreemiumVisit Runloop (official site)
6. 

   SPACE

   Open-source sandbox platform from Perplexity that runs agent and user code in isolated Firecracker microVMs through an HTTP API.

   Developer platformsCodingAutomation and agents
   Open sourceVisit SPACE (official site)

E2B 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 docs.e2b.dev · fetched 23 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.