E2BSandbox integration

Integrate with the E2BSandbox sandbox using LangChain Python.

E2B provides cloud sandboxes for running AI-generated code in isolated environments. See the E2B docs for signup, authentication, and platform details.

Installation

pip

pip install langchain-e2b

uv add langchain-e2b

Authentication

Set your E2B API key:

export E2B_API_KEY="..."

Configure E2B-specific options, such as templates or sandbox lifetime, through the E2B SDK when creating the sandbox.

Create a sandbox

In Python, you create the sandbox using the E2B SDK, then wrap it with the deepagents backend.

from e2b import Sandbox
from langchain_e2b import E2BSandbox

e2b_sandbox = Sandbox.create()
backend = E2BSandbox(sandbox=e2b_sandbox)

try:
    result = backend.execute("echo hello")
    print(result.output)
finally:
    e2b_sandbox.kill()

Use with Deep Agents

from e2b import Sandbox
from deepagents import create_deep_agent
from langchain_anthropic import ChatAnthropic
from langchain_e2b import E2BSandbox

e2b_sandbox = Sandbox.create()
backend = E2BSandbox(sandbox=e2b_sandbox)

agent = create_deep_agent(
    model=ChatAnthropic(model="claude-sonnet-4-6"),
    system_prompt="You are a coding assistant with sandbox access.",
    backend=backend,
)

try:
    result = agent.invoke(
        {
            "messages": [
                {"role": "user", "content": "Create a hello world Python script and run it"}
            ]
        }
    )
finally:
    e2b_sandbox.kill()

Use with Deep Agents Code

langchain-e2b also publishes an E2B sandbox provider for Deep Agents Code, so dcode can run agent tool calls inside an E2B sandbox. Install the package into the dcode environment, set your API key, then select the provider:

Using E2B with Deep Agents Code requires dcode 0.1.19 or newer and langchain-e2b 0.0.4 or newer.

dcode --install langchain-e2b --package
export E2B_API_KEY="..."
dcode --sandbox e2b

E2B sandboxes use /home/user as the default working directory. To create sandboxes from a specific E2B template, set E2B_TEMPLATE; to change the sandbox lifetime, set E2B_SANDBOX_TIMEOUT (in seconds). Each variable also accepts a DEEPAGENTS_CODE_-prefixed form (for example, DEEPAGENTS_CODE_E2B_API_KEY), which takes precedence.

Deep Agents Code manages the sandbox lifecycle for you, creating a sandbox on start and deleting it on exit.

Cleanup

Always kill E2B sandboxes when you are done to avoid ongoing resource usage.

See also: Sandboxes.

---

Connect these docs to Claude, VSCode, and more via MCP for real-time answers.

Edit this page on GitHub or file an issue.