Overview

Updated at:

Code Interpreter runs code inside a sandbox and returns structured results. It is suitable for data analysis, notebook-style experiences, immediate execution of agent-generated code, and returning tables, text, errors, and file artifacts to a business system.

SDK entry points

Code Interpreter uses a dedicated SDK package:

import { Sandbox } from "@e2b/code-interpreter";

In Python:

from e2b_code_interpreter import Sandbox

Connection parameters still use E2B_API_KEY, E2B_API_URL, and E2B_DOMAIN. FC Agent Sandbox provides the built-in Code Interpreter template code-interpreter-v1, and examples should specify that template explicitly.

Minimal example

import os

from e2b_code_interpreter import Sandbox

sandbox = Sandbox.create(
    template="code-interpreter-v1",
    api_key=os.environ["E2B_API_KEY"],
    api_url=os.environ["E2B_API_URL"],
    domain=os.environ["E2B_DOMAIN"],
)

try:
    execution = sandbox.run_code("print('hello')")
    print("".join(execution.logs.stdout or []).strip())
finally:
    sandbox.kill()

Core objects

  • sandbox.run_code() / sandbox.runCode(): run code and return the execution result.

  • execution.logs.stdout / execution.logs.stderr: standard output and standard error.

  • execution.text: the text result of the final bare expression.

  • execution.results: a rich-result list that can include text, tables, and other structured results.

  • execution.error: runtime errors, including error type, value, and traceback.

  • Code Context: preserves variables, imports, and the working directory across multiple executions.

Further reading