Welcome to Orichain’s documentation!

Orichain (/ori’chain/) is a lightweight, Python-based library designed specifically for Retrieval-Augmented Generation (RAG) use cases. Built for seamless integration with your endpoints, Orichain simplifies the process of writing, maintaining, and reviewing RAG workflows.

While inspired by LangChain, Orichain focuses on performance and efficiency. Its fully asynchronous and threaded architecture ensures high concurrency and responsiveness out-of-the-box - so you can focus on building, not optimizing.

Note

• The Library was rewritten in v2.0.0 released in March of 2025. Starting with this version, each module now provides separate synchronous and asynchronous classes, giving you greater flexibility to choose the approach that best fits your application.

• Starting from version 2.0.1 (May 2025), a provider argument is now required when initializing the LLM and EmbeddingModel classes.

---

Get started here:

Contents:

• Installation
• Getting Started
  • Creating Embeddings with Orichain
  • Generating LLM Responses with Orichain
  • Retrieving Documents with Orichain Knowledge Base
  • Detecting Language with Orichain
  • Next Steps
• Modules
  • orichain.embeddings
  • orichain.llm
  • orichain.knowledge_base
  • orichain.lang_detect
• Code Examples
  • Tool Calling with orichain.llm
  • Using Orichain in a Production-Ready FastAPI Application

---

💡 Additional Resources

• 📂 Source Code: Visit the Orichain GitHub Repository for reporting issues, exploring the codebase, or contributing to the project.