Hi all - so my goal is to basically build an iPhone app using a ChatGPT backed character, which users can interact with by speaking (speech to text) and then will hear a spoken reply (text to speech)
I'll need to use APIs that allow commercial usage.
I'm trying to wrap my head around the costs of such a project. Right now I assume I'll have API costs from
1.) Speech to text (like whisper API)
2.) LLM (ChatGPT API)
3.) Text to speech (say elevenlabs API)
If a ton of people start using this app, how fast am I going broke lol?
I figure I can give free usage up to a point, and then users can pay for additional use if they like the service.
But what do you guys recommend as the most cost effective way to do this? Looking at Elevenlabs alone, that looks like it would become super expensive very quickly.
Any other APIs that allow commercial products which you would recommend?
Or does this project sound like a fools errand?
Any input would be greatly appreciated! Thank you!
[link] [comments]