feat(dictation): expose auth_time_ms in DictationResponse

CLAUDE.md

@@ -51,7 +51,7 @@ aai.settings.api_key = "your-key"
 - `aai.DictationTranscriber` — Dictation API: audio uploaded as it is spoken over one live request, transcript (and optional LLM pass) out. Methods: `open_live()` (push-style `DictationLiveSession`: `write()`/`close()`/`result()`/`abort()`), `transcribe_live()` (iterable, file object, bytes or path), `warm()`. There is no buffered `transcribe()`: every entry point posts to `/v1/transcribe/live`
 - `aai.AsyncDictationTranscriber` — Asyncio counterpart of `DictationTranscriber`; `open_live()` returns an `AsyncDictationLiveSession`. Owns an HTTP pool: use `async with` or `await aclose()`
 - `aai.DictationConfig` — Dictation options: `sample_rate`, `channels`, `language_codes`, `keyterms_prompt`, `llm_instruction`. Rejects unknown fields
-- `aai.DictationResponse` — Dictation result: `.text`, `.words` (`DictationWord` with `text` + `confidence`), `.confidence`, `.llm_response`, `.llm_error`, `.audio_duration_ms`, `.session_id`, `.request_time_ms`, `.sync_time_ms`, and `.final_text` (the LLM rewrite, falling back to `.text`)
+- `aai.DictationResponse` — Dictation result: `.text`, `.words` (`DictationWord` with `text` + `confidence`), `.confidence`, `.llm_response`, `.llm_error`, `.audio_duration_ms`, `.session_id`, `.request_time_ms`, `.sync_time_ms`, `.auth_time_ms`, and `.final_text` (the LLM rewrite, falling back to `.text`)
 - `assemblyai.streaming.v3.RealTimeTranscriber` — Real-time streaming with event-based API (threaded)
 - `assemblyai.streaming.v3.AsyncRealTimeTranscriber` — Asyncio-native counterpart; same options/events
 - `aai.LLMGateway` — LLM Gateway client. Resources: `models.list()`, `chat.completions.create()` (incl. `stream=True`), `understanding.create()`/`.validate()`

assemblyai/__version__.py

@@ -1 +1 @@
-__version__ = "1.6.1"
+__version__ = "1.6.2"

assemblyai/dictation/v1/models.py

@@ -192,6 +192,10 @@ class DictationResponse(BaseModel):
     sync_time_ms: Optional[float] = None
     "Time in milliseconds spent transcribing, excluding the LLM pass."
 
+    auth_time_ms: Optional[float] = None
+    """Authentication and key-validation time in milliseconds, as a portion
+    of request_time_ms. ``None`` when the server predates the field."""
+
     @property
     def final_text(self) -> str:
         """

tests/unit/test_dictation.py

@@ -43,6 +43,7 @@
     "session_id": "eb92c4ff-4bbb-429f-9b99-7279d7fe738f",
     "request_time_ms": 243.7,
     "sync_time_ms": 180.2,
+    "auth_time_ms": 24.6,
 }
 
 
@@ -288,6 +289,24 @@ def test_transcribe_live_parses_response(httpx_mock: HTTPXMock):
     assert result.audio_duration_ms == 400
     assert result.request_time_ms == 243.7
     assert result.sync_time_ms == 180.2
+    assert result.auth_time_ms == 24.6
+
+
+def test_transcribe_live_parses_response_without_auth_time(httpx_mock: HTTPXMock):
+    # Given a server response that predates the auth_time_ms field
+    response = {k: v for k, v in _OK_RESPONSE.items() if k != "auth_time_ms"}
+    httpx_mock.add_response(
+        url=LIVE_URL,
+        method="POST",
+        status_code=httpx.codes.OK,
+        json=response,
+    )
+
+    # When streaming audio chunks
+    result = aai.DictationTranscriber().transcribe_live(_chunks(b"RIFF", b"fake"))
+
+    # Then auth_time_ms is None instead of a parse failure
+    assert result.auth_time_ms is None
 
 
 def test_transcribe_live_sends_raw_api_key_to_the_dictation_host(httpx_mock: HTTPXMock):

tests/unit/test_dictation_async.py

@@ -39,6 +39,7 @@
     "session_id": "eb92c4ff-4bbb-429f-9b99-7279d7fe738f",
     "request_time_ms": 243.7,
     "sync_time_ms": 180.2,
+    "auth_time_ms": 24.6,
 }
 
 
@@ -195,6 +196,25 @@ async def test_transcribe_live_parses_response(httpx_mock: HTTPXMock):
     assert result.words[1].text == "reports"
     assert result.request_time_ms == 243.7
     assert result.sync_time_ms == 180.2
+    assert result.auth_time_ms == 24.6
+
+
+async def test_transcribe_live_parses_response_without_auth_time(httpx_mock: HTTPXMock):
+    # Given a server response that predates the auth_time_ms field
+    response = {k: v for k, v in _OK_RESPONSE.items() if k != "auth_time_ms"}
+    httpx_mock.add_response(
+        url=LIVE_URL,
+        method="POST",
+        status_code=httpx.codes.OK,
+        json=response,
+    )
+
+    # When streaming from an async producer
+    async with aai.AsyncDictationTranscriber() as transcriber:
+        result = await transcriber.transcribe_live(_achunks(b"RIFF", b"fake"))
+
+    # Then auth_time_ms is None instead of a parse failure
+    assert result.auth_time_ms is None
 
 
 async def test_transcribe_live_posts_to_the_dictation_host(httpx_mock: HTTPXMock):