-
Notifications
You must be signed in to change notification settings - Fork 2k
Description
I would like to propose adding support for Replicate, which is a provider for various AI Models for text generation, image/video/audio, embeddings and more.
Replicate provides access to those models through a unified API where only Input/Output Values differ depending on the selected model.
I already have the code changes ready and use it a lot in my personal projects.
It would be an honour for me to contribute back to spring-ai.
Implementation / Expected Behaviour
I would suggest an implementation that provides four model classes to handle various response types:
- ReplicateChatModel (implements ChatModel)
Handles text-based LLMS like GPT, Claude, Llama, DeepSeek etc. - ReplicateMediaModel
Handles image, video and audio generation models like Flux, Sora etc. - ReplicateStructuredModel
Handles models returning structured JSON objects like embeddings, OCR etc. - ReplicateStringModel
Handles Models that return simple String outputs like Content Classifiers.
Current Behavior
Currently there is no support for Replicate.com.
Future Idea
Replicate also supports training, fine-tuning and deploying custom models. We could also add support for this in the future as those opens up a lot of interesting use cases.
I have a PR ready and wanted to ask if this is something you think would be valuable for spring ai. I would be more than happy to contribute