This repository implements a chatbot designed for question answering over the LangChain documentation. It targets developers and users seeking to query technical documentation efficiently, leveraging real-time updates and streaming capabilities.

How It Works

The system comprises two main components: ingestion and question-answering. Ingestion involves fetching HTML from the LangChain documentation site and GitHub codebase, loading it with RecursiveURLLoader and SitemapLoader, splitting documents using RecursiveCharacterTextSplitter, and creating a Weaviate vectorstore with OpenAI embeddings. The question-answering pipeline first refines user input and chat history into a standalone question using an LLM, retrieves relevant documents from the vectorstore, and then passes these to a model for a streamed answer.

Quick Start & Requirements

This project is deployed via LangGraph Cloud and is not designed for local execution without a LangGraph Cloud account. For a local, albeit feature-limited, version, refer to the code and documentation from the specified branch.

Highlighted Details

• Built using LangChain, LangGraph, and Next.js.
• Leverages LangGraph's streaming support and async API for real-time updates.
• Includes comprehensive documentation on concepts, modification, LangSmith integration, production readiness, and deployment.
• Features a trace URL generation for chat sessions and an endpoint for feedback collection.

Maintenance & Community

No specific details on contributors, sponsorships, or community channels (like Discord/Slack) are provided in the README.

Licensing & Compatibility

The repository's licensing is not specified in the provided README.

Limitations & Caveats

The project is currently deployed via LangGraph Cloud, limiting local execution without an account. A separate branch offers a local-runnable version but with a reduced feature set.