Bug Description
I'm trying to run Amazon Nova Sonic realtime model. I get this error
Traceback (most recent call last):
File "/usr/local/lib/python3.13/dist-packages/livekit/plugins/aws/__init__.py", line 32, in __getattr__
from .experimental import realtime
File "/usr/local/lib/python3.13/dist-packages/livekit/plugins/aws/experimental/realtime/__init__.py", line 1, in <module>
from .realtime_model import RealtimeModel, RealtimeSession
File "/usr/local/lib/python3.13/dist-packages/livekit/plugins/aws/experimental/realtime/realtime_model.py", line 18, in <module>
from aws_sdk_bedrock_runtime.client import (
...<2 lines>...
)
ImportError: cannot import name 'BedrockRuntimeClient' from 'aws_sdk_bedrock_runtime.client' (/usr/local/lib/python3.13/dist-packages/aws_sdk_bedrock_runtime/client.py)
ImportError: The 'realtime' module requires optional dependencies. Please install them with: pip install 'livekit-plugins-aws[realtime]'
Yet I have already installed 'livekit-plugins-aws[realtime]'
Expected Behavior
The nova sonic model should work with this error and the agent should start speaking. All the other realtime models work fine.
Reproduction Steps
Steps to reproduce the behavior:
1. Run the Voice AI Quickstart using Amazon Nova Sonic Realtime model
2. You will get the above error
Operating System
macOs
Models Used
Amazon Nova Sonic RealtimeModel
Package Versions
- Python 3.14
- livekit 1.1.15
- livekit-agents 1.7.1
- livekit-plugins-aws 1.7.1
Session/Room/Call IDs
No response
Proposed Solution
Additional Context
No response
Screenshots and Recordings
Here is a screenshot of my pyproject.toml file

Bug Description
I'm trying to run Amazon Nova Sonic realtime model. I get this error
Yet I have already installed
'livekit-plugins-aws[realtime]'Expected Behavior
The nova sonic model should work with this error and the agent should start speaking. All the other realtime models work fine.
Reproduction Steps
Operating System
macOs
Models Used
Amazon Nova Sonic RealtimeModel
Package Versions
Session/Room/Call IDs
No response
Proposed Solution
Additional Context
No response
Screenshots and Recordings
Here is a screenshot of my
pyproject.tomlfile