chore: sync sdk code with DeepLearning repo (#186)

Co-authored-by: AssemblyAI <engineering.sdk@assemblyai.com>

Author: MAsuamah

README.md

@@ -18,9 +18,6 @@
 
 With a single API call, get access to AI models built on the latest AI breakthroughs to transcribe and understand audio and speech data securely at large scale.
 
-> **⚠️ WARNING**
-> This SDK is intended for **testing and light usage only**. It is not recommended for use at scale or with production traffic. For best results, we recommend calling the AssemblyAI API directly via HTTP request. See our [official documentation](https://www.assemblyai.com/docs) for more information, including HTTP code examples.
-
 # Overview
 
 - [AssemblyAI's Python SDK](#assemblyais-python-sdk)

assemblyai/types.py

@@ -2358,6 +2358,9 @@ class BaseTranscript(BaseModel):
     temperature: Optional[float] = None
     "Change how deterministic the response is, with 0 being the most deterministic and 1 being the least deterministic."
 
+    remove_audio_tags: Optional[str] = None
+    "When set to 'all', removes all bracketed audio/speaker tags (e.g. [MUSIC], [Speaker: A]) from the transcript. Only supported for Universal-3 Pro."
+
     keyterms_prompt: Optional[List[str]] = None
     "The list of key terms used to generate the transcript with the Slam-1 speech model. Can't be used together with `prompt`."