trigger-realtime

Real-time task monitoring and streaming for Trigger.dev tasks across frontend and backend. Subscribe to task runs and stream data in real-time from frontend and backend. See references/realtime.md for complete documentation. - Subscribe to task runs with status updates, progress tracking, and output streaming via backend iterators or React hooks - Stream AI/LLM responses directly to React components using typed stream definitions that pipe OpenAI completions and other async iterables - Implement human-in-the-loop workflows with wait tokens that pause task execution until frontend approval or user input is received - Secure frontend access through scoped public tokens with fine-grained permissions (read-only, task-specific, time-limited) and trigger tokens for task initiation - Building progress indicators for long-running tasks - Creating live dashboards showing task status - Streaming AI/LLM responses to the UI - React components that trigger and monitor tasks - Waiting for user approval in tasks - Scope tokens narrowly — only grant necessary permissions - Set expiration times — don't use long-lived tokens - Use typed hooks — pass task types for proper inference - Handle errors — always check for errors in hooks - Throttle streams — use throttleInMs to control re-renders