<https://blog.livekit.io/xai-livekit-partnership-g...
# show-and-tell
b
Very excited about the xAI Grok Voice Agent API available in LiveKit. Couple of questions: - What's the context window limit for this model? Couldn't find it anywhere. GPT-Realtime-Mini is 32K and wondering if this is higher, would love to move off of there. - In the xAI docs, it mentions this API has access built-in to searching the web, searching X and searching collections we upload. Is this supported in LiveKit already or will it be coming soon? Thanks!
p
Pretty neat! Any idea when this will be available for NodeJS users?
r
I am not sure when it will be available in NodeJS. You can learn more about the details of Grok here: https://docs.x.ai/docs/overview
b
Does LiveKit have access to the above capabilities? Is it built in? The documentation isn't very clear nor does it mention context window length
r
Looks like context window size is on the first page there. It says 2M
It has Features • Function calling • Structured outputs • Reasoning • Lightning fast • Low cost Using it, it seems to be sourcing content from search
But I am not 100% sure about the search feature.
You would need to talk to the xAI team about that. We are providing the connectors
👍 1
You can try it here: https://grok.livekit.io/
b
Thanks
m
Could this be used with avatar plugins to reduce latency?
r
Model:
grok-4-1-fast-non-reasoning
From their tweet they bill @ $0.05 / minute for the realtime but be sure to verify that.