Fix NVIDIA SageMaker region config

CHANGELOG.md

@@ -5,6 +5,13 @@ All notable changes to **Pipecat** will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [1.2.1] - 2026-05-15
+
+### Changed
+
+- Updated NVIDIA SageMaker STT and TTS generated service configs to use
+  `AWS_REGION` for the SageMaker endpoint region.
+
 ## [1.2.0] - 2026-05-15
 
 ### Added

src/pipecat_cli/registry/_configs.py

@@ -72,9 +72,9 @@
     "nvidia_stt": 'NvidiaSTTService(api_key=os.getenv("NVIDIA_API_KEY"))',
     "nvidia_sagemaker_stt": (
         "NvidiaSageMakerSTTService(\n"
-        '        endpoint_name=os.getenv("NVIDIA_SAGEMAKER_STT_ENDPOINT_NAME"),\n'
-        '        region=os.getenv("NVIDIA_SAGEMAKER_STT_REGION")\n'
-        "    )\n"
+        '    endpoint_name=os.getenv("NVIDIA_SAGEMAKER_STT_ENDPOINT_NAME"),\n'
+        '    region=os.getenv("AWS_REGION")\n'
+        ")\n"
     ),
     "openai_stt": 'OpenAISTTService(api_key=os.getenv("OPENAI_API_KEY"))',
     "openai_realtime_stt": 'OpenAIRealtimeSTTService(api_key=os.getenv("OPENAI_API_KEY"))',
@@ -471,12 +471,12 @@
     ),
     "nvidia_sagemaker_tts": (
         "NvidiaSageMakerTTSService(\n"
-        '        endpoint_name=os.getenv("NVIDIA_SAGEMAKER_TTS_ENDPOINT_NAME"),\n'
-        '        region=os.getenv("NVIDIA_SAGEMAKER_TTS_REGION"),\n'
-        "        settings=NvidiaSageMakerTTSService.Settings(\n"
-        '            voice=os.getenv("NVIDIA_SAGEMAKER_TTS_VOICE_ID"),\n'
-        "        ),\n"
-        "    )\n"
+        '    endpoint_name=os.getenv("NVIDIA_SAGEMAKER_TTS_ENDPOINT_NAME"),\n'
+        '    region=os.getenv("AWS_REGION"),\n'
+        "    settings=NvidiaSageMakerTTSService.Settings(\n"
+        '        voice=os.getenv("NVIDIA_SAGEMAKER_TTS_VOICE_ID"),\n'
+        "    ),\n"
+        ")\n"
     ),
     "openai_tts": (
         "OpenAITTSService(\n"

src/pipecat_cli/registry/service_metadata.py

@@ -394,7 +394,7 @@ class ServiceRegistry:
         ServiceDefinition(
             value="nvidia_sagemaker_stt",
             label="NVIDIA SageMaker",
-            package="pipecat-ai[nvidia,sagemaker]",
+            package="pipecat-ai[aws,sagemaker]",
             class_name=["NvidiaSageMakerSTTService"],
             env_prefix="NVIDIA_SAGEMAKER_STT",
             include_params=["endpoint_name", "region"],
@@ -856,7 +856,7 @@ class ServiceRegistry:
         ServiceDefinition(
             value="nvidia_sagemaker_tts",
             label="NVIDIA SageMaker",
-            package="pipecat-ai[nvidia,sagemaker]",
+            package="pipecat-ai[aws,sagemaker]",
             class_name=["NvidiaSageMakerTTSService"],
             env_prefix="NVIDIA_SAGEMAKER_TTS",
             include_params=["endpoint_name", "region"],
@@ -1062,6 +1062,21 @@ class ServiceRegistry:
 # These services have complex initialization logic that cannot be auto-generated
 # (e.g., nested InputParams, SessionProperties, or other special requirements)
 MANUAL_SERVICE_CONFIGS = {
+    "nvidia_sagemaker_stt": (
+        "NvidiaSageMakerSTTService(\n"
+        '    endpoint_name=os.getenv("NVIDIA_SAGEMAKER_STT_ENDPOINT_NAME"),\n'
+        '    region=os.getenv("AWS_REGION")\n'
+        ")"
+    ),
+    "nvidia_sagemaker_tts": (
+        "NvidiaSageMakerTTSService(\n"
+        '    endpoint_name=os.getenv("NVIDIA_SAGEMAKER_TTS_ENDPOINT_NAME"),\n'
+        '    region=os.getenv("AWS_REGION"),\n'
+        "    settings=NvidiaSageMakerTTSService.Settings(\n"
+        '        voice=os.getenv("NVIDIA_SAGEMAKER_TTS_VOICE_ID"),\n'
+        "    ),\n"
+        ")"
+    ),
     "aws_bedrock_llm": (
         "AWSBedrockLLMService(\n"
         '    aws_region=os.getenv("AWS_REGION"),\n'

src/pipecat_cli/templates/server/env.example.jinja2

@@ -57,8 +57,8 @@ PLIVO_AUTH_TOKEN=
 ASSEMBLYAI_API_KEY=
 
 {% endif %}
-{% if stt_service == 'aws_transcribe_stt' or llm_service == 'aws_bedrock_llm' or tts_service == 'aws_polly_tts' or realtime_service == 'aws_nova_realtime' %}
-# AWS (STT/LLM/TTS/Realtime)
+{% if stt_service == 'aws_transcribe_stt' or stt_service == 'nvidia_sagemaker_stt' or llm_service == 'aws_bedrock_llm' or tts_service == 'aws_polly_tts' or tts_service == 'nvidia_sagemaker_tts' or realtime_service == 'aws_nova_realtime' %}
+# AWS (STT/LLM/TTS/Realtime/SageMaker)
 AWS_ACCESS_KEY_ID=
 AWS_SECRET_ACCESS_KEY=
 AWS_REGION=
@@ -201,13 +201,11 @@ NVIDIA_MODEL=
 {% if stt_service == 'nvidia_sagemaker_stt' %}
 # NVIDIA SageMaker (STT)
 NVIDIA_SAGEMAKER_STT_ENDPOINT_NAME=
-NVIDIA_SAGEMAKER_STT_REGION=
 
 {% endif %}
 {% if tts_service == 'nvidia_sagemaker_tts' %}
 # NVIDIA SageMaker (TTS)
 NVIDIA_SAGEMAKER_TTS_ENDPOINT_NAME=
-NVIDIA_SAGEMAKER_TTS_REGION=
 NVIDIA_SAGEMAKER_TTS_VOICE_ID= # "Magpie-Multilingual.EN-US.Isabella"
 
 {% endif %}

tests/test_service_registry.py

@@ -110,6 +110,21 @@ def test_get_service_config(self):
         assert "DeepgramSTTService" in config
         assert "DEEPGRAM_API_KEY" in config
 
+    def test_nvidia_sagemaker_configs_use_aws_region(self):
+        """Test NVIDIA SageMaker services use shared AWS_REGION."""
+        stt_config = ServiceLoader.get_service_config("nvidia_sagemaker_stt")
+        tts_config = ServiceLoader.get_service_config("nvidia_sagemaker_tts")
+
+        assert stt_config is not None
+        assert tts_config is not None
+        assert 'endpoint_name=os.getenv("NVIDIA_SAGEMAKER_STT_ENDPOINT_NAME")' in stt_config
+        assert 'endpoint_name=os.getenv("NVIDIA_SAGEMAKER_TTS_ENDPOINT_NAME")' in tts_config
+        assert 'region=os.getenv("AWS_REGION")' in stt_config
+        assert 'region=os.getenv("AWS_REGION")' in tts_config
+        assert "NVIDIA_SAGEMAKER_STT_REGION" not in stt_config
+        assert "NVIDIA_SAGEMAKER_TTS_REGION" not in tts_config
+        assert 'voice=os.getenv("NVIDIA_SAGEMAKER_TTS_VOICE_ID")' in tts_config
+
     def test_get_service_import(self):
         """Test retrieving service imports."""
         imports = ServiceLoader.get_service_import("deepgram_stt")