[FEATURE] Support for Replicate #410
Description
Scope check
- This is core LLM communication (not application logic)
- This benefits most users (not just my use case)
- This can't be solved in application code with current RubyLLM
- I read the Contributing Guide
Due diligence
- I searched existing issues
- I checked the documentation
What problem does this solve?
Adds Replicate to the list of providers. Replicate is an inference provider, it gives access to many models. Not just text or image generation, also video, speech, music, upscaling images... It also allows users to fine-tune and host own models.
There are other well known inference providers out there: Fal.ai, Groq, Hugging Face also has inference, etc. Supporting one would probably open the door to supporting others.
Proposed solution
Adding support for Replicate.
Maybe it'd require adding much more functionality to support different model capabilities.
Or maybe initially we'd filter out the models that are not for text or image. I'm not sure there's a straightforward way to do that via Replicate API. I also didn't see how to get the different model pricing's from their API, I understand these are desireable things.
Why this belongs in RubyLLM
Support for Replicate will allow supporting many more models, of different capabilities. And even if we stick with only text and image, it's still adding more models, and more options for whoever is thinking on using RubyLLM.
I think supporting this opens the door to supporting other inference providers, and there are a few.