Skip to main content
Source: src/cwsandbox/_function.py:40 
class RemoteFunction(fn: Callable[P, R], *, session: Session, container_image: str | None = None, serialization: Serialization = Serialization.JSON, temp_dir: str = DEFAULT_TEMP_DIR, profile_ids: list[str] | None = None, profile_names: list[str] | None = None, runner_ids: list[str] | None = None, resources: ResourceOptions | dict[str, Any] | None = None, mounted_files: list[dict[str, Any]] | None = None, s3_mount: dict[str, Any] | None = None, ports: list[dict[str, Any]] | None = None, network: NetworkOptions | dict[str, Any] | None = None, max_timeout_seconds: int | None = None, environment_variables: dict[str, str] | None = None, annotations: dict[str, str] | None = None)
Wrapper for remote function execution in sandboxes. RemoteFunction wraps a Python function for execution in a sandbox. It provides methods to execute remotely, locally (for testing), and map over multiple inputs in parallel. The wrapped function must be synchronous. Async functions are not supported.

​
Methods

​
remote

remote(*args: P.args = (), **kwargs: P.kwargs = {}) -> OperationRef[R]
Execute the function in a sandbox, return OperationRef immediately. Parameters
  • *args (P.args): Positional arguments to pass to the function.
  • **kwargs (P.kwargs): Keyword arguments to pass to the function.
Returns
  • OperationRef[R]: OperationRef[R]: Use .result() to block until result is ready.
Examples 
ref = compute.remote(2, 3)
result = ref.result()  # Block for result
# Or in async context:
result = await ref

​
map

map(items: Iterable[tuple[Any, ...]]) -> list[OperationRef[R]]
Execute the function for each item, return OperationRefs immediately. Each item should be a tuple of positional arguments for the function. All executions are launched in parallel. Parameters
  • items (Iterable[tuple[Any, ...]]): Iterable of argument tuples to pass to the function.
Returns
  • list[OperationRef[R]]: List of OperationRef[R], one for each item.
Examples 
# Execute add(1, 2), add(3, 4), add(5, 6) in parallel
refs = add.map([(1, 2), (3, 4), (5, 6)])
results = [ref.result() for ref in refs]  # [3, 7, 11]

​
local

local(*args: P.args = (), **kwargs: P.kwargs = {}) -> R
Execute the function locally without a sandbox. Useful for testing and debugging. Runs the original function directly in the current Python process. Parameters
  • *args (P.args): Positional arguments to pass to the function.
  • **kwargs (P.kwargs): Keyword arguments to pass to the function.
Returns
  • R: The result of the function execution.
Examples 
# Test without sandbox overhead
result = compute.local(2, 3)
assert result == 5
Last modified on May 19, 2026