invoke-ai/InvokeAI · error · ValueError

A saved workflow must be selected before executing call_save

Error message

A saved workflow must be selected before executing call_saved_workflow.

What it means

call_saved_workflow refuses to run when no workflow_id is set. The node requires a workflow that has been saved to the workflow records service; validate_selected_workflow checks this before anything else and raises ValueError immediately if the field is empty/None.

Source

Thrown at invokeai/app/invocations/call_saved_workflow.py:53

    classification=Classification.Beta,
)
class CallSavedWorkflowInvocation(BaseInvocation):
    """Displays and later executes against a selected saved workflow."""

    workflow_id: str = InputField(
        default="",
        description="The selected saved workflow ID, managed by the workflow editor UI.",
        ui_type=UIType.SavedWorkflow,
    )
    workflow_inputs: dict[str, Any] = InputField(
        default={},
        description="Literal values for the selected workflow's exposed inputs, managed by the workflow editor UI.",
        ui_hidden=True,
    )

    def validate_selected_workflow(self, context: InvocationContext):
        if not self.workflow_id:
            raise ValueError("A saved workflow must be selected before executing call_saved_workflow.")

        try:
            workflow_record = context._services.workflow_records.get(self.workflow_id)
        except WorkflowNotFoundError as e:
            raise ValueError(f"The selected saved workflow '{self.workflow_id}' could not be found.") from e

        config = context._services.configuration
        if config.multiuser:
            queue_user_id = context._data.queue_item.user_id
            user = context._services.users.get(queue_user_id)
            is_admin = bool(user and user.is_admin)
            is_owner = workflow_record.user_id == queue_user_id
            is_default = workflow_record.workflow.meta.category is WorkflowCategory.Default
            if not (is_default or is_owner or workflow_record.is_public or is_admin):
                raise ValueError(f"The selected saved workflow '{self.workflow_id}' is not accessible to this user.")

        return workflow_record

View on GitHub (pinned to 0b6a024f2f)

Solutions

  1. Select a saved workflow for the node (set workflow_id to the saved workflow's ID) before executing.
  2. Programmatically set workflow_id on the CallSavedWorkflow node before queueing the graph.
  3. Add a pre-flight check on the graph to ensure every call_saved_workflow node has a non-empty workflow_id.

Example fix

// before
node = CallSavedWorkflow()  # workflow_id not set
graph.add_node(node)
// after
node = CallSavedWorkflow(workflow_id="saved-workflow-uuid")
graph.add_node(node)
Defensive patterns

Strategy: validation

Validate before calling

if not node.workflow_id:
    raise ValueError("select a saved workflow before executing")

Try / catch

try:
    output = node.invoke(context)
except ValueError as e:
    if "must be selected" in str(e):
        prompt_user_to_select_workflow()
    else:
        raise

Prevention

When it happens

Trigger: Invoking a CallSavedWorkflow invocation (directly via invoke() or through run_node) while self.workflow_id is empty string or None — e.g. the node was added to a graph but the user never picked a workflow in the editor, or the field was programmatically left at its default.

Common situations: Building graphs via API/script where the workflow_id field was never populated; UI workflows where the user forgot to select a saved workflow; importing a graph that references a node whose workflow_id was never persisted.

Understand the failure class

Background: "Missing required field" and "field is required" errors: why libraries reject payloads that omit mandatory fields — this error's family across 20 libraries.

Related errors

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/7a1514c95cc1a915. Report an issue: GitHub.