dev.gxuri.in

Fri Sep 12 2025

Kolophon - No-Code RAG Application Builder

Kolophon - No-Code RAG Application Builder

Next.js

LangChain

Pinecone

ElevenLabs

Drizzle ORM

PostgreSQL

Kolophon is a no-code platform for building Retrieval-Augmented Generation (RAG) applications such as AI chatbots, voice assistants, and analytics dashboards using your own data.

What is Kolophon?


Introduction

Kolophon is a no-code platform for building Retrieval-Augmented Generation (RAG) applications that allows anyone to create AI-powered chatbots, voice assistants, and analytics dashboards using their own data.

Instead of writing complex backend pipelines, users can upload datasets, configure prompts, and instantly deploy an intelligent AI interface powered by modern large language models.

The platform integrates LangChain for RAG pipelines, Pinecone for vector search, Groq and Google GenAI for LLM inference, and ElevenLabs for voice synthesis. Built using Next.js App Router, Kolophon provides a scalable architecture for deploying AI applications with minimal configuration.

Users can create bots, embed them into websites, analyze interaction metrics, and manage datasets all within a single dashboard.

Features

  1. No-Code RAG Builder

    • Upload documents or datasets.
    • Automatically generate embeddings.
    • Query data through AI-powered chat interfaces.

  2. Multiple AI Interfaces

    • Chat-based conversational agents.
    • Voice-enabled assistants with real-time recording and speech synthesis.
    • Embeddable widgets for websites.

  3. Dataset-Powered Intelligence

    • Uses vector embeddings stored in Pinecone.
    • Retrieves relevant context before sending prompts to LLMs.
    • Ensures answers are grounded in user-provided data.

  4. Interactive Analytics Dashboard

    • Tracks bot interactions across daily, weekly, and monthly intervals.
    • Visualized using Recharts charts.
    • Helps measure engagement and usage patterns.

  5. Embeddable AI Widgets

    • Chat and voice widgets can be embedded into external websites.
    • Supports iframe-friendly routes for easy integration.

  6. Authentication & User Management

    • Secure login using NextAuth.
    • Supports providers like Google and GitHub.

  7. Scalable Database Architecture

    • PostgreSQL database managed via Drizzle ORM.
    • Stores users, projects, and analytics data.

  8. Modern UI System

    • Built with Tailwind CSS and Shadcn/UI components.
    • Smooth interactions powered by Lenis scrolling.
    • Modular component-based architecture.

User Workflow

  1. Users sign in using their preferred authentication provider.
  2. A new AI project (bot) is created through the dashboard.
  3. Users upload datasets that are converted into embeddings.
  4. Kolophon stores vectors in Pinecone for semantic retrieval.
  5. When a query is sent to the bot:
    • Relevant documents are retrieved.
    • Context is passed to the LLM through LangChain.
  6. The AI returns a grounded response based on the uploaded data.
  7. User interactions are logged and visualized in the analytics dashboard.
  8. The bot can be embedded into external websites via chat or voice widgets.

Tech Stack

  • Next.js 15 (App Router): Full-stack framework for frontend and backend routes.
  • React 19: UI framework for building interactive interfaces.
  • LangChain: RAG orchestration and LLM pipelines.
  • Groq / Google GenAI: LLM inference providers.
  • Pinecone: Vector database for semantic search.
  • ElevenLabs: Text-to-speech and voice interaction.
  • PostgreSQL: Primary database.
  • Drizzle ORM: Type-safe database access layer.
  • NextAuth: Authentication and session management.
  • Tailwind CSS + Shadcn UI: Component-driven design system.
  • Recharts: Data visualization for analytics.

How it Works

If you want to explore the implementation details, start with these key areas of the codebase:

  • /api/bot/query — main RAG query endpoint
  • /api/bot/create — project creation logic
  • /api/bot/analytics — analytics data retrieval
  • /embed/ — embeddable chat and voice widgets
  • /src/components/ — reusable UI components

These files demonstrate how Kolophon connects dataset embeddings, LLM inference, and analytics tracking into a unified platform.

Visit the GitHub repository to explore the full implementation, contribute improvements, or build your own AI applications on top of the platform.

Stay Connected

Follow me for updates on this project and other AI applications: