Awesome Infra for AI › Prompt Management

patterns-ai-core/langchainrb

⭐ 2000 Ruby repository created 2023-04-18

Langchain.rb is a Ruby library designed for developing applications powered by Large Language Models (LLMs). It offers a consistent, unified interface to interact with a multitude of LLM providers including OpenAI, Anthropic, Google Gemini, AWS Bedrock, and others, abstracting away provider-specific API differences. The library facilitates core LLM operations such as generating text completions, chat completions, and embeddings, enabling developers to easily switch between different LLM backends. Key features include robust prompt management through prompt templates (including few-shot templates), output parsers for structuring LLM responses, and tools for building Retrieval Augmented Generation (RAG) systems. It explicitly supports vector search, a crucial component for RAG, by integrating with various vector database technologies. Langchain.rb also includes functionalities for evaluation ("evals") to assess model performance and supports various logging mechanisms. While it focuses on the application development and operational aspects of LLMs rather than model training, it serves as a critical component for deploying and operating AI agent applications and LLM workflows in a Ruby environment. Its primary purpose is to simplify the operationalization of LLM-based features within applications.

https://github.com/patterns-ai-core/langchainrb

rubylangchainllmai-agentsprompt-managementragvector-searchllm-gatewaymodel-serving