RubyLLM: A Ruby framework for all major AI providers
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- RubyLLM launches as a unified Ruby framework offering a single interface across 14+ AI providers: OpenAI, xAI, Anthropic, Gemini, VertexAI, Bedrock, DeepSeek, Mistral, Ollama, OpenRouter, Perplexity, GPUStack, and any OpenAI-compatible API
- The framework ships with only three dependencies — Faraday, Zeitwerk, and Marcel — and ships a registry of 800+ models with automatic capability detection and pricing baked in
- RubyLLM consolidates chat, vision (images and video), audio transcription, document extraction (PDF, CSV, JSON), image generation, embeddings, and content moderation into one library, alongside a tool-calling DSL and a
RubyLLM::Agentclass for reusable assistants - The library supports structured JSON output via a schema DSL, streaming responses via Ruby blocks, fiber-based async concurrency, and persistent "extended thinking" reasoning traces from supported models
- Rails integration comes via the
acts_as_chatActiveRecord mixin and four generators —ruby_llm:install,db:migrate,ruby_llm:load_models, andruby_llm:chat_ui— that scaffold a working browser-based chat app at/chatswith a singlebin/rails generatecommand
Why it matters: Ruby teams building AI features currently stitch together vendor-specific SDKs with mismatched response shapes; RubyLLM replaces that with one gem, three dependencies, and 14+ providers, so swapping from GPT-5-nano to Claude Sonnet 4 or a local Ollama model is a config change rather than a rewrite. The Rails-native persistence layer and bundled chat-UI generator compress the path from `bundle install` to a production-shaped chatbot into minutes.
Ask SkimNews
