RubyLLM
A Ruby-native AI framework, for plain Ruby or Rails, offering chats, tools, agents, image and video generation, embeddings, reranking, transcription, speech and OCR through one consistent API across 19 model providers.
RubyLLM lets a Ruby or Rails application work with models, tools and agents through one consistent API. Chats can ask questions, attach files such as images, video, audio or PDFs, stream responses chunk by chunk, and call registered tools; the same API also generates images and videos, creates and reranks embeddings, transcribes audio to text, turns text into speech, extracts document text as markdown through OCR, and checks whether content should be moderated.
It supports 19 model providers and lets an application move between hosted and local providers, or connect an OpenAI-compatible endpoint, without rewriting application code. Version 2.0 is the current release line, with the previous 1.x series still documented separately.