StopReason

type StopReason =
  | "cancelled"
  | "contentFiltered"
  | "endTurn"
  | "guardrailIntervened"
  | "interrupt"
  | "maxTokens"
  | "limitOutputTokens"
  | "limitTotalTokens"
  | "limitTurns"
  | "pauseTurn"
  | "refusal"
  | "stopSequence"
  | "toolUse"
  | "modelContextWindowExceeded"
  | string & {
};

Defined in: src/types/messages.ts:670

Reason why the model stopped generating content.

• cancelled - Agent invocation was cancelled via agent.cancel()
• contentFiltered - Content was filtered by safety mechanisms
• endTurn - Natural end of the model’s turn
• guardrailIntervened - A guardrail policy stopped generation
• interrupt - Agent execution was interrupted for human input
• maxTokens - The model provider’s per-call token cap was reached
• limitOutputTokens - Agent loop stopped because InvokeOptions.limits.outputTokens was reached
• limitTotalTokens - Agent loop stopped because InvokeOptions.limits.totalTokens was reached
• limitTurns - Agent loop stopped because InvokeOptions.limits.turns was reached
• pauseTurn - Model paused a long-running turn; the response should be sent back to continue
• refusal - A streaming classifier intervened to handle a potential policy violation
• stopSequence - A stop sequence was encountered
• toolUse - Model wants to use a tool
• modelContextWindowExceeded - Input exceeded the model’s context window