LangChain Daytona

0.0.5 verified Fri May 01 auth: no python

LangChain Daytona provides a Dayntona sandbox integration for secure code execution within LangChain agents. Version 0.0.5 requires Python >=3.11 and <4.0. Part of the deepagents monorepo, it is under active development.

pip install langchain-daytona

Common errors

error ModuleNotFoundError: No module named 'langchain_daytona' ↓

cause Package not installed or installed under a different name.

fix

Run pip install langchain-daytona (note the hyphen, not underscore).

error AttributeError: module 'langchain_daytona' has no attribute 'DaytonaSandboxTool' ↓

cause The symbol may be in a submodule or the package version is too old.

fix

Ensure langchain-daytona>=0.0.5. Then import as from langchain_daytona import DaytonaSandboxTool. If still missing, check if the package exposes a different name.

error ValueError: DAYTONA_API_KEY environment variable not set ↓

cause The required API key is missing.

fix

Set the DAYTONA_API_KEY environment variable before running your script, e.g., export DAYTONA_API_KEY='your_key'.

Warnings

breaking Version 0.0.5 drops support for Python <3.11. Upgrade Python or pin to an older version. ↓

fix Use Python 3.11 or later, or install langchain-daytona==0.0.4 if you need older Python.

gotcha The DaytonaSandboxTool requires a running Daytona sandbox instance. Without it, invocations will fail with a connection error. ↓

fix Ensure DAYTONA_API_KEY environment variable is set and the sandbox ID is valid and active.

deprecated The old import path langchain_daytona.tools.DaytonaSandboxTool is deprecated and will be removed in a future release. ↓

fix Use from langchain_daytona import DaytonaSandboxTool instead.

Imports

DaytonaSandboxTool
wrong
```
from langchain_daytona.tools import DaytonaSandboxTool
```
correct
```
from langchain_daytona import DaytonaSandboxTool
```
Tool is re-exported at package level; submodule import may work but not guaranteed.

Quickstart

Initialize a Daytona sandbox tool with an API key and sandbox ID, then run code inside the sandbox.

import os
from langchain_daytona import DaytonaSandboxTool

tool = DaytonaSandboxTool(api_key=os.environ.get('DAYTONA_API_KEY', ''), sandbox_id="my-sandbox")
result = tool.invoke({"code": "print('hello from sandbox')"})
print(result)