A full-stack Retrieval-Augmented Generation (RAG) chatbot built with Next.js, AI SDK, Drizzle ORM, and Neon Postgres. It allows users to chat with AI that can retrieve information from a custom knowledge base.
- Frontend: React, Next.js 14 (App Router), Tailwind CSS, AI Elements
- Backend: Next.js API routes, AI SDK (for embeddings, retrieval, and streaming)
- Database: Neon (PostgreSQL with pgvector)
- ORM: Drizzle ORM
- Auth: Clerk
- Deployment: None
- Retrieval-Augmented Generation (RAG) pipeline
- Upload and chunk custom data sources
- Embeddings generation and vector search with Neon
- Streaming AI chat responses
- Authentication and user sessions via Clerk
- Full-stack TypeScript setup with Drizzle ORM
- Clean and responsive chat UI built with AI Elements
Home page
Chatbot UI
Click here for detailed documentation
This RAG Chatbot can be applied in a variety of real-world scenarios where context-aware and domain-specific answers are required:
- Customer Support Automation
- Integrate with your website or product to provide instant, accurate answers to frequently asked questions using your internal knowledge base.
- Internal Knowledge Management
- Employees can query company documents, policies, or technical manuals quickly, reducing time spent searching for information.
- Education & Training
- Students or trainees can ask questions about course materials, manuals, or technical content and get precise, sourced responses.
- Content Retrieval for SaaS or Enterprise Apps
- Enhance your product by enabling users to interact with large sets of documentation, wikis, or reports through a chat interface.
- Personal Knowledge Assistant
- Users can feed in personal notes, research papers, or PDFs to create a smart assistant that answers questions based on their own data.
Essentially, any scenario where users need AI to provide accurate answers grounded in a custom dataset or documents can benefit from this chatbot.
Built with ❤️ by @nikitapoyarekar05