# Livekit inference on self hosted

**URL:** <https://community.livekit.io/t/livekit-inference-on-self-hosted/1743>\
**Category:** Getting Started\
**Tags:** agent-development, livekit-inference, livekit-cloud\
**Created:** [July 27, 2026, 1:59pm UTC](https://community.livekit.io/t/livekit-inference-on-self-hosted/1743 "2026-07-27T13:59:25Z")\
**Posts on this page:** 1\
**Showing post:** 2

<div class="post-metadata">

**Author:** ![darryncampbell](https://sea1.discourse-cdn.com/flex001/user_avatar/community.livekit.io/darryncampbell/32/20_2.png) [@darryncampbell](https://community.livekit.io/u/darryncampbell)\
**Post date:** [July 27, 2026, 10:40pm UTC](https://community.livekit.io/t/livekit-inference-on-self-hosted/1743/2 "2026-07-27T22:40:04Z")

</div>

Just to clarify, I’m interpreting your question as **self-hosted LiveKit server**. You can call Inference models through self-hosted Agents using LiveKit cloud as your transport without issue.

Technically, yes, it does work, but it’s not officially recommended or ‘supported’. There was an interesting thread recently with a user who had done just this, but they were seeing lower TTFT and missed caching compared with agents running on cloud, so you may find you wouldn’t get the benefits you’re looking for anyway:

> [@Gemma-4-31B TTFT on shared gateway (~400ms floor) vs self-hosted (28ms), same checkpoint](https://community.livekit.io/t/gemma-4-31b-ttft-on-shared-gateway-400ms-floor-vs-self-hosted-28ms-same-checkpoint/1704):
>
> Hi, We’re building voice agents on livekit-agents (1.6.5) with google/gemma-4-31b-it via LiveKit Inference. The launch post advertises 192ms TTFT, but on the shared gateway (agent-gateway.livekit.cloud) we’re seeing about a 400ms floor, rising toward a second at realistic voice context sizes. To figure out how much is the model and how much is the serving path, we served the identical checkpoint ourselves on one H100 with stock vLLM. Numbers, recipe, and questions are below. Happy to share all …

---

_[View the full topic](https://community.livekit.io/t/livekit-inference-on-self-hosted/1743)._
