can you say more about that please? the
# office-hours
e
can you say more about that please? the call cost will report Vapi + transcription cost, but I'm using my own LLM and 11labs keys
s
We only charge you for the providers for whom you are using our API keys. So, if you are using your own keys for the LLM and TTS providers, then in the final cost, their cost will be zero.
call_id please?
e
yes thats I mean - can Vapi report anything to do with LLM or 11labs use if its my keys? e.g. approximate tokens or approximate number of words etc
@Shubham Bajaj I checked, and it looks like promptTokens etc are just reported as "0" when I use my own API keys:
Copy code
"costBreakdown": {
    "transport": 0,
    "stt": 0.0073,
    "llm": 0,
    "tts": 0,
    "vapi": 0.0661,
    "chat": 0,
    "total": 0.0734,
    "llmPromptTokens": 0,
    "llmCompletionTokens": 0,
    "ttsCharacters": 0,
Can you suggest any way in which I can calculate call costs when I'm using my own LLM and 11labs API keys? Can VAPI still provide promptTokens, ttsTokens etc when using my keys? Without that, it's very difficult to know the cost of calls.
s
Actually, we don't keep the logs for usage and cost when you use your own keys. You have to check with your LLM and text-to-speech provider if they can help you with this either through some tags or something else by which you can track the calls through them.
e
I see - any reason why not? You would still be able to get the same metrics when you call the LLM/TTS correct? It's pretty difficult for us to link it when you make the calls
2 Views