https://livekit.io logo
Join Slack
Powered by
MiniMax TTS model missing language parameter in Py...
# ask-ai
n
MiniMax TTS model missing language parameter in Python SDK
t
Thanks for reaching out, Yuhan. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.
Thanks — I’m looking into the missing 
language
parameter for the MiniMax TTS in the Python SDK. I can help but need one quick clarification so I give the right guidance: Which integration are you using? • the MiniMax plugin directly via 
livekit-agents[minimax]
, or • the LiveKit Inference layer using 
inference.TTS
in an 
AgentSession
? Also please share: • the 
livekit-agents
package version (
pip show livekit-agents
), and • a short code snippet showing how you’re calling the TTS API (or the exact error you see). Quick notes while you fetch that: • MiniMax’s docs do not list 
language
as a top-level parameter. Language can often be passed via 
extra_kwargs
or controlled by model selection. • Using 
inference.TTS
lets you pass provider-specific options through 
extra_kwargs
if you need language control. Send the answers/snippet and I’ll give exact code you can try. Sources: MiniMax TTS plugin guide | LiveKit docs | Cartesia TTS | LiveKit docs | Deepgram TTS | LiveKit docs | Rime TTS | LiveKit docs
n
I'm using livekit.plugins.minimax.TTS, it does not support language_boost, which is supported in native minimax api
t
Thanks — I hear you’re trying to use MiniMax’s native 
language_boost
from 
livekit.plugins.minimax.TTS
and it’s not available. That’s correct: the MiniMax plugin does not expose 
language_boost
as a top-level parameter. What you can do • Pass provider-specific options (like 
language_boost
) via 
extra_kwargs
if you’re using the plugin entry that accepts it. • Or use the 
inference.TTS
path in an 
AgentSession
, which explicitly forwards provider-specific options through 
extra_kwargs
. Example (plugin + 
extra_kwargs
):
Copy code
from livekit.plugins.minimax import TTS

tts = TTS(
    model="your-model",
    voice="your-voice",
    extra_kwargs={"language_boost": 1.5}
)
audio = tts.synthesize("Text to speak")
Example (inference.TTS via AgentSession):
Copy code
from livekit.inference import TTS as InferenceTTS

inference_tts = InferenceTTS(
    model="minimax",
    voice="your-voice",
    extra_kwargs={"language_boost": 1.5}
)
# attach to AgentSession as your TTS provider
Next steps I can help with • Share your 
livekit-agents
package version (
pip show livekit-agents
) so I can confirm if a newer release added direct support. • If 
extra_kwargs
doesn’t work in your setup, share a short snippet of how you instantiate and call 
TTS
so I can suggest an exact fix.
n
I'm using Version: 1.3.5
t
Thanks — I see you’re on v1.3.5. That version does not expose 
language_boost
as a top-level argument on 
livekit.plugins.minimax.TTS
. You can still use the native MiniMax option by passing it through 
extra_kwargs
, or by using 
inference.TTS
in an 
AgentSession
, which also forwards provider-specific options. Quick examples: 1) Using the MiniMax plugin with 
extra_kwargs
Copy code
from livekit.plugins.minimax import TTS

tts = TTS(
    model="your-model",
    voice="your-voice",
    extra_kwargs={"language_boost": 1.5}
)
audio = tts.synthesize("Text to speak")
2) Using 
inference.TTS
(AgentSession)
Copy code
from livekit.inference import TTS as InferenceTTS

inference_tts = InferenceTTS(
    model="minimax",
    voice="your-voice",
    extra_kwargs={"language_boost": 1.5}
)
# attach inference_tts to your AgentSession as the TTS provider
If 
extra_kwargs
doesn’t take effect in your setup, please paste the small snippet where you instantiate and call 
TTS
(and any error/output). I’ll suggest the exact fix or confirm whether upgrading the package is needed.
n
TypeError: TTS.__init__() got an unexpected keyword argument 'extra_kwargs'
t
Thanks — I see the TypeError and what you’re trying to do: you attempted to pass 
language_boost
via 
extra_kwargs
to 
livekit.plugins.minimax.TTS
, but that class does not accept 
extra_kwargs
and does not expose 
language_boost
. That’s why you get the TypeError. What you can do next: • Use the Inference path that does accept provider options: create an 
inference.TTS
instance and attach it to your `AgentSession`; 
inference.TTS
accepts 
extra_kwargs
and will forward 
language_boost
. • Or call the MiniMax HTTP API directly from your code if 
language_boost
is essential and you need immediate control. • Optionally check for a newer plugin release, but v1.3.5’s 
livekit.plugins.minimax.TTS
does not support 
language_boost
. If you want, paste the minimal snippet that shows how you currently instantiate and call 
minimax.TTS
and I’ll provide the exact code to switch to 
inference.TTS
or an HTTP call example. Sources: MiniMax TTS plugin guide | LiveKit docs
2 Views
Open in Slack
PreviousNext
Powered by