i am testing my phone agent. i made one call after...
# agents
a
i am testing my phone agent. i made one call after another and now i see "Concurrent STT is 20", going to exceed my plan limit. Since i made calls one after another, why the concurrent STT became 20? Does only inference STT count into concurrent STT? STT solution in plugins doesn't count, right? I created STT plugin with Qwen APIs.
đź‘€ 1
r
What agent library are you using and what version?
f
Cc: @astonishing-accountant-25155 Hey Franklin, Thanks for posting in Livekit community. “Concurrent STT” is about how many STT streams/connections are open at the same time (usually WebSockets), not how many calls you made total...
With phone agents, each call typically opens an STT stream and keeps it open for the whole AgentSession. If previous sessions aren’t fully closing (or take a bit to close), the concurrency can climb even if you dial “one after another”.
Also: if your code is still using LiveKit Inference for STT (e.g., inference.STT(...) or an STT model string), that does count toward the “Concurrent STT” limit. A custom plugin that calls Qwenn directly shouldn’t count against LiveKit Inference STT, but it only works if you’re actually passing that plugin into AgentSession as stt=...