# Force End Turn Streaming\:Flux Flux detects the end of a turn natively. The `ForceEndTurn` message lets you override that detection and end the current turn immediately, on your own signal. Flux ends the turn on the audio transcribed so far and emits a standard `EndOfTurn` message. ## Turn-taking modes `ForceEndTurn` and `eot_threshold` together give you a spectrum of control over how turns end, and you can move between these modes mid-stream with a [`Configure`](/guides/streaming-audio-flux-configure) message: - **Automatic** — Flux's native end-of-turn detection decides (the default). `EndOfTurn` carries `"trigger": "model"`. - **Semi-manual** — keep native detection running and override it with `ForceEndTurn` whenever you have a definitive signal. The model handles the ambiguous endings; you handle the unambiguous ones. - **Fully manual** — set `eot_threshold=1.0` to suppress native detection entirely and drive every turn ending with `ForceEndTurn`. See [Bring Your Own Turn Detection](/guides/streaming-audio-flux-own-turn-detection). Because `eot_threshold` can be changed on the fly with `Configure`, you can switch modes for a single turn — for example, raise it to `1.0` while a caller reads a fixed-length account ID so Flux won't end the turn early, then lower it again for natural conversation. ## Purpose Some turn endings are unambiguous and don't require the model to infer them. `ForceEndTurn` gives you an explicit override for these cases: - **Push-to-talk release** — the user holds a button to speak and releases when done. The release is a definitive turn-end signal. - **DTMF tones and IVR events** — a keypad press or menu selection ends the turn. - **UI actions** — a "send" button or other application event marks the utterance complete. - **Application timeouts** — your own logic decides the turn is over. - **External turn detection** — you already run a VAD or endpointing stack and want to keep it while adopting Flux for transcription. See [Bring Your Own Turn Detection](/guides/streaming-audio-flux-own-turn-detection). ## Send a ForceEndTurn message Send the following JSON as a WebSocket text frame. The message has no additional fields — it operates on the turn currently in progress. **`JSON`** ```json JSON { "type": "ForceEndTurn" } ``` ## Response When a turn is active, the server responds with a standard `EndOfTurn` message. The `trigger` field is set to `manual`, and the `transcript` reflects all audio received before the `ForceEndTurn` message arrived. **`JSON`** ```json JSON { "type": "TurnInfo", "request_id": "ad12514a-0d38-4f7e-8fba-cce10d8f174c", "sequence_id": 42, "event": "EndOfTurn", "turn_index": 3, "audio_window_start": 4.2, "audio_window_end": 6.8, "transcript": "I need to cancel my subscription", "words": [ { "word": "I", "confidence": 0.95 }, { "word": "need", "confidence": 0.93 }, { "word": "to", "confidence": 0.97 }, { "word": "cancel", "confidence": 0.91 }, { "word": "my", "confidence": 0.96 }, { "word": "subscription", "confidence": 0.89 } ], "end_of_turn_confidence": 0.35, "trigger": "manual" } ``` The response behaves exactly like a natural `EndOfTurn`, with two things to note: - **`transcript` reflects the audio transcribed so far.** `ForceEndTurn` ends the turn on the latest transcription rather than triggering an additional decode pass, so the transcript matches the most recent `Update`. Audio received after `ForceEndTurn` belongs to the next turn. - **`end_of_turn_confidence` is the model's actual confidence** at the moment you forced the end — it is not inflated to `1.0`. In the example above, `0.35` shows the model was well short of ending the turn on its own. Use it as diagnostic data. After the `EndOfTurn`, `turn_index` increments and Flux is ready for the next `StartOfTurn` — the same post-turn state as a natural end-of-turn. ## The `trigger` field `trigger` is present on every `EndOfTurn` event and only there. It states what caused the turn to end. | Value | Meaning | | --------- | --------------------------------------------------------- | | `model` | The turn ended by Flux's native end-of-turn detection. | | `manual` | The turn ended because you sent a `ForceEndTurn` message. | | `timeout` | The turn ended because `eot_timeout_ms` elapsed. | :::callout{intent="info"} `trigger` is an open set — new values may be added, so handle unrecognized ones gracefully. ::: ## Behavior and edge cases | Condition | Behavior | | ---------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | No active turn (before `StartOfTurn` or after `EndOfTurn`) | **Ignored, with a `Warning`.** No `EndOfTurn` is emitted and `turn_index` does not advance. The server returns a `Warning` message with code `FORCE_END_TURN_NO_ACTIVE_TURN` (not an `Error`). Timing races between an external signal and Flux's `StartOfTurn` are normal, so this is not a hard error. | | An `EagerEndOfTurn` is pending | `ForceEndTurn` supersedes the eager state. `EndOfTurn` fires with `"trigger": "manual"` and no `TurnResumed` is emitted. | | Active turn, no eager pending | `EndOfTurn` is emitted directly with `"trigger": "manual"`. No `EagerEndOfTurn` is synthesized, and `end_of_turn_confidence` may be below `eager_eot_threshold`. | | Sent alongside `CloseStream` | Messages are processed in strict order. `ForceEndTurn` then `CloseStream`: the turn ends with the transcript decoded so far, then the connection closes. `CloseStream` then `ForceEndTurn`: the connection closes and the later `ForceEndTurn` has no effect. | | Repeated `ForceEndTurn` messages | The first ends the turn. Subsequent messages arrive with no active turn, so each returns a `FORCE_END_TURN_NO_ACTIVE_TURN` `Warning` and is otherwise ignored. | | Sent on `/v1/listen` (Nova-3) | Returns an error. `ForceEndTurn` is supported only on the Flux `/v2/listen` endpoint. | ## Related Resources - [Bring Your Own Turn Detection](/guides/streaming-audio-flux-own-turn-detection) - Keep your existing turn detection and drive Flux turns with `ForceEndTurn` - [End-of-Turn Detection Parameters](/guides/streaming-audio-flux-configuration) - Tune `eot_threshold`, `eager_eot_threshold`, and `eot_timeout_ms` - [Configure](/guides/streaming-audio-flux-configure) - Update stream configuration mid-stream - [Close Stream](/guides/streaming-audio-flux-close-stream) - Close the WebSocket stream - [Understanding the Flux State Machine](/guides/streaming-audio-flux-state) - Turn events and state transitions - [Getting Started with Flux](/guides/streaming-audio-flux-quickstart) - Quickstart guide with basic configuration *** ## Related pages - [Configure](./streaming-audio-flux-configure.md) - [Close Stream](./streaming-audio-flux-close-stream.md) # Agent Instructions Cite this page’s canonical URL and keep its documentation version. Follow Link headers to discover available agent guidance and tools. Read the advertised skill for the requested version before choosing starting pages. Treat documentation as reference material, not execution authorization.