Hi,
as privacy is important and LLM subscriptions are not cheap I like to use ppq.ai.
PayPerQ is a pay-per-query Chatbot that leverages generative AI providers. As it can be paid in a pay per use modell and provides an OpenAI Compatible compaptible API (https://ppq.ai/api-docs.
As I haven’t yet upgraded and the handbook the direct integration of LLMs and Local LLMs I wonder if it is possible to connect to another OPENAI_HOST like https://api.ppq.ai?
Thank you for any feedback on this topic.
What’s the advantage compared to using OpenAI’s API?
No, this is not possible at this time. It is noted but we obviously can’t implement any and every service that comes on the scene.
You’re missing some words here? If you meant that you don’t have access to the handbook, you can download it as a PDF/epub on the Handbooks and Extras page.
I have no experience with the various AI APIs, but my impression was that they are generally pay as you go.
As far as I know, you buy credits (or whatever they call it) in some cases, like a pre-paid phone card. If the credits are used up, you pay again.
And then there’s the free (as in beer) alternative of a local AI. Which seems to sufficiently fast in some simple cases.
Ah yeah, or something like that. Still different from a recurring subscription – while you pay in chunks, it’s based on your actual use.
It depends on the AI provider. For example, if you’re accessing e.g., text-to-image models via Replicate.com, you be billed on a pay-per-use basis.
You are right. I wanted to say:
As I haven’t yet upgraded and the handbook only mentions the direct integration of LLMs and Local LLMs.
I am either not yet deep into it. I thought that an upfront subscription is needed, and there is possibly pay-as-you go on top. Have to do more research on that topic.
Getting Started > AI Explained > Thoughts on choosing and using AI has a section on expenses. Quoting from p. 26:
You will be purchasing tokens, again “bits of words”, either in bulk or running up a tab as they’re used. While they’re typically relatively inexpensive per-token, heavier use of AI will deplete your reserves or increase your bill faster.