> Fetch clean Markdown by appending `.md` to any page URL under https://signalwire.com/docs or requesting it with the HTTP header `Accept: text/markdown`. The root index at https://signalwire.com/docs/llms.txt lists the available documentation indexes.

# detect_answering_machine

> Detect whether a human or an answering machine answered a call.

[detect]: /docs/server-sdks/reference/python/relay/call/detect

Detect a human versus an answering machine. A typed convenience over
[`detect()`][detect] that builds the `{"type": "machine", "params": {...}}`
configuration for you. All parameters are keyword-only, and only the ones you
set are sent.

## Parameters

**`initial_timeout`** `float | None` — default: None

Seconds to wait for initial voice before deciding.

---

**`end_silence_timeout`** `float | None` — default: None

Seconds of silence that end the greeting.

---

**`machine_voice_threshold`** `float | None` — default: None

Seconds of continuous voice above which the greeting is treated as a machine.

---

**`machine_words_threshold`** `int | None` — default: None

Word count above which the greeting is treated as a machine.

---

**`detect_interruptions`** `bool | None` — default: None

When `True`, detect the caller interrupting during the machine greeting
playback.

---

**`detect_message_end`** `bool | None` — default: None

Whether to keep listening for the end of a machine's message, so you can
leave one after the beep.

---

**`timeout`** `float | None` — default: None

Maximum seconds to run the detector before stopping. Keyword-only.

---

**`on_completed`** `Callable[[RelayEvent], Any] | None` — default: None

Callback invoked when the operation reaches a terminal state. Can be a regular
function or async coroutine. Keyword-only.

---

## Returns

[`DetectAction`](/docs/server-sdks/reference/python/relay/actions) -- An action handle with `stop()` and `wait()` methods. It resolves on the first detection result.

## Example

```python {12-15}
from signalwire.relay import RelayClient

client = RelayClient(
    project="your-project-id",
    token="your-api-token",
    contexts=["default"],
)

@client.on_call
async def handle_call(call):
    await call.answer()
    action = await call.detect_answering_machine(
        detect_message_end=True,
        timeout=30,
    )
    event = await action.wait()
    print("Detect result:", event.params)

client.run()
```

> Detect whether a human or an answering machine answered a call.