Can someone explain billing. 1. The Call billing breakdown which has LLM, STT and TTS, is this final costs or i am getting double charged by openai, anthropic for using their APIs ? 2. I wondering how should i interpret billing, should i fund my account in openai and anthropic or should i only fund vapi ? Cost STT: $0.39 ($0.03 / min) LLM: $0.92 ($0.06 / min) - 85462 prompt tokens - 2045 completion tokens TTS: $0.13 ($0.01 / min) - 8711 characters Vapi: $0.75 ($0.05 / min) Transport: $0.00 ($0.00 / min) Total: $2.19 ($0.15 / min)

Nope if you use their APIs, you won't be charged here through Vapi

HI @jordan I am seeing this on my dashboard. I use my own keys for 11 labs. Is this plot for what users usually pay when theu use VAPI for all these services? I am really interested in knowing 0.04 /minute for 11 labs. Sounds too good!

These are the averages we see on the platform, might not be 100% accurate! Best thing to do is to try it yourself and check the cost breakdown in the calls log

Thanks @jordan so, is the LLM cost I see an estimate of what could it be ? but when I look at the main stripe billing page, it seems like it’s the addition of all calls. Something seems off.

Yes that's just a rough estimate based on the model and prompt length. You'll want to use the cost breakdown in the calls page to get a real average

hi @jordan I meant this one, i am seeing LLM cost even though i am using my LLM key. And i believe in the main billing section, the overall cost: Total: $2.19 is getting charged to vapi account Cost STT: $0.39 ($0.03 / min) LLM: $0.92 ($0.06 / min) 85462 prompt tokens 2045 completion tokens TTS: $0.13 ($0.01 / min) 8711 characters Vapi: $0.75 ($0.05 / min) Transport: $0.00 ($0.00 / min) Total: $2.19 ($0.15 / min)

Yes I believe this is because we fall back to our latency optimized deployments when OpenAI is seeing high traffic to keep latency low. It will be at-cost though, same as OpenAI.