how does the cost of the avatar calculated when using livekit inferernce

Thanks for reaching out, Mohammed. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.

Running the Node starter repo uses LiveKit Cloud for media/agent hosting. If your agent is configured to call STT/LLM/TTS via the

inference

module, you are also using LiveKit Inference (billed separately). Self‑hosting means you run your own agent/server infrastructure and manage model provider keys yourself unless you choose to route model calls through LiveKit Inference. does that mean its billing both livekit inference and livekitcloud? how do i change it to not use livekit cloud? does that require self-hosting?

Thanks — I’m responding to your questions about whether you’ll be billed for both LiveKit Cloud and LiveKit Inference, and how to stop using LiveKit Cloud. Short answers • Yes: if you run the Node starter repo configured to use LiveKit Cloud for media/agent hosting and your agent calls STT/LLM/TTS via the

inference

module, you will be billed for both LiveKit Cloud (media/transport) and LiveKit Inference (model usage) separately. • To avoid LiveKit Cloud for media, you must self‑host your LiveKit server (run your own LiveKit instance) so media transport no longer uses LiveKit Cloud. • To avoid LiveKit Inference billing, do not route STT/LLM/TTS calls through the

inference

module — instead, have your self‑hosted agent call your chosen model provider(s) directly using your own provider keys. What each billing line covers • LiveKit Cloud: charges for real‑time media transport — WebRTC minutes, bandwidth/data transfer, and any plan fees. • LiveKit Inference: charges for STT, LLM, and TTS model usage when you route those model calls through LiveKit’s inference service. • Avatar/provider rendering: billed separately by the avatar provider (if used). How to stop using LiveKit Cloud (high level) 1. Self‑host LiveKit server: ◦ Run your own LiveKit instance instead of using LiveKit Cloud. That moves media/agent hosting off LiveKit Cloud so you no longer incur LiveKit Cloud transport charges (you’ll keep infrastructure costs for the VM/container/network where you host LiveKit). 1. Manage inference calls yourself: ◦ Configure your agent to call STT/LLM/TTS providers directly using your provider API keys rather than using the

inference

module that routes through LiveKit Inference. 1. Update the Node starter repo configuration: ◦ Point the client and server to your self‑hosted LiveKit server URL and use the appropriate API/key settings. ◦ Replace or remove the

inference

integration so model calls go to your chosen providers. Notes and tradeoffs • Self‑hosting removes LiveKit Cloud transport billing but does not eliminate infrastructure costs (compute, egress, maintenance). • If you self‑host but still use LiveKit Inference for STT/LLM/TTS, you will still be billed for inference. • If you self‑host and also self‑host model calls (or use third‑party model providers directly), you will avoid LiveKit Inference charges but must manage provider keys, scaling, and reliability yourself. If you want, I can help with next steps: • Walk through specific config changes in the Node starter repo to point to a self‑hosted LiveKit server. • Show how to remove/replace the

inference

module calls so your agent calls model providers directly. • Help estimate costs comparing LiveKit Cloud + Inference vs self‑hosting (you’ll need to share expected session length, participant counts, and which model providers you plan to use). Which of those would you like to do next?

help me with this Help estimate costs comparing LiveKit Cloud + Inference vs self‑hosting (you’ll need to share expected session length, participant counts, and which model providers you plan to use). u need to 3 diff calculations livekit cloud transport + livekit inference() livekit self hosting + livekit inference livekit self hosting + livekit plugins also give there latency and setup difficulty difference models: gpt 5 mini deepgram nova 3 multi inworld tts 1.5 max anam avatar (12 dollar starter pack per month, 0.16 dollars overages)

I don't have the answer you're looking for. You could also try asking your question: • in one of the other Slack channels or • to https://deepwiki.com/livekit/livekit_composite which is trained on all LiveKit source code If you find the answer, please post it here to help others!

if i want to use livekit inference $5 in credits (~100 minutes) then billed based on model prices des that mean i have to buy the ship pacakage?