Documentation Index
Fetch the complete documentation index at: https://docs.coreweave.com/llms.txt
Use this file to discover all available pages before exploring further.
This guide covers how to use environment variables in sandboxes.
Environment variables should not be used for sensitive information like API keys, passwords, or other secrets.
For sensitive values, use the Secret type instead. Secrets are resolved server-side and never sent as raw values by the client. Basic usage
Set environment variables when creating a sandbox:
from cwsandbox import Sandbox
with Sandbox.run(
environment_variables={"LOG_LEVEL": "info"},
) as sandbox:
result = sandbox.exec([
"python",
"-c",
"import os; print(os.environ.get('LOG_LEVEL'));",
]).result()
print(result.stdout.strip()) # "info"
Session-level defaults
Use sessions to share environment variables across multiple sandboxes:
from cwsandbox import SandboxDefaults, Session
defaults = SandboxDefaults(
environment_variables={
"PROJECT_ID": "my-project",
"LOG_LEVEL": "info",
},
)
with Session(defaults) as session:
with session.sandbox() as sb1:
result = sb1.exec([
"python",
"-c",
"import os; print(os.environ.get('LOG_LEVEL'));",
]).result()
print(result.stdout.strip()) # "info"
# Override LOG_LEVEL and add new variable
with session.sandbox(
environment_variables={
"LOG_LEVEL": "debug", # Override session default
"MODEL_NAME": "gpt-4", # Add new variable
}
) as sb2:
result = sb2.exec([
"python",
"-c",
"import os; "
"print(os.environ.get('PROJECT_ID')); "
"print(os.environ.get('LOG_LEVEL')); "
"print(os.environ.get('MODEL_NAME'));",
]).result()
lines = result.stdout.strip().split("\n")
print(lines) # ["my-project", "debug", "gpt-4"]
Remote functions
Environment variables work with remote functions:
with Session(defaults) as session:
@session.function(environment_variables={"MODEL_VERSION": "v2.0"})
def process(task_id: int) -> dict:
import os
return {
"task": task_id,
"project": os.environ.get("PROJECT_ID"), # From session defaults
"version": os.environ.get("MODEL_VERSION"), # From function decorator
}
result = process.remote(42).result()
print(result) # {"task": 42, "project": "my-project", "version": "v2.0"}
Environment variables are passed by reference. Mutations will be reflected in subsequent function calls:
env_vars = {"MODEL_VERSION": "v2.0"}
@session.function(environment_variables=env_vars)
def process(task_id: int) -> dict:
import os
return {"version": os.environ.get("MODEL_VERSION")}
result = process.remote(42).result() # version: "v2.0"
env_vars["MODEL_VERSION"] = "v3.0" # Mutate the dictionary
result = process.remote(42).result() # version: "v3.0" (changed)
