Fast transformer inference for Ruby

State-of-the-art transformers for Ruby

LLM Memory is a Ruby gem designed to provide large language models (LLMs) like ChatGPT with memory using in-context learning. This enables better integration with systems such as Rails and web services while providing a more user-friendly and abstract interface based on brain terms.

OpenAI CLIP embeddings, uses ONNX models. Allows to create embeddings for images and text

A Ruby gem providing a consistent interface for various AI/ML tokenizers including OpenAI GPT, Anthropic Claude, Google Gemini, Meta Llama, Mistral, Qwen, and embedding models like BERT, BGE, and multilingual-E5. Features caching, truncation, token counting, and error handling across different tokenization libraries.

A Ruby port of FastEmbed - fast text embeddings using ONNX Runtime

A Ruby-native interface for fine-tuning machine learning models, starting with image classification using SigLIP2

Basic Ruby inference wrapper for the GLiNER2 ONNX model.

Easy to use some onnx models.

Easy to use some onnx models.

A minimal, hackable LLM implementation with inference capabilities. Port of Karpathy's nanochat.

Security pipelines for LLM applications using ONNX models from Hugging Face Hub. Detect prompt injections, jailbreaks, and PII leaks. Models are lazily downloaded and cached locally. Fast local inference (~10-20ms after initial load).

Ruby implementation of Google's SigLIP2 model for creating text and image embeddings. Uses ONNX models from HuggingFace onnx-community.

TokenEstimator is a Rails gem that allows you to count tokens in Excel, CSV, PDF, TXT, Markdown, and input text files using different tokenizers.

Vector Search backed by your vector search DB of choice.