---

🤖 Overview

AI ChatKit is a production-ready, full-stack AI chat framework that lets you drop a fully featured chat interface into any application in minutes. It abstracts away provider-specific API quirks — streaming protocols, token limits, tool-call formats — behind a single clean interface so your team ships features instead of glue code.

Problem it solves: Every team building with LLMs rewrites the same boilerplate — provider adapters, SSE streaming, conversation history, auth, and embeddings. AI ChatKit is that solved layer, plug-and-play out of the box.

---

✨ Features

Feature	Description
🔌 Multi-Provider LLM Support	Switch between OpenAI, Google Gemini, Anthropic Claude, DeepSeek, and Qwen with a single env var
⚡ Streaming Responses	Real-time token streaming via Server-Sent Events (SSE) — zero polling
🧠 Conversation History	Persistent, multi-turn conversation storage with SQLite or PostgreSQL
🏢 Multi-Tenant Architecture	Isolated sessions and API keys per workspace or user
🔗 REST API	Clean, documented REST endpoints — integrate with any frontend or service
🪄 Embeddable Widget	Drop the React component into any existing app with a single import
📚 RAG-Ready	ChromaDB vector store integration with bge-m3 multilingual embeddings
🔒 JWT Auth	Stateless authentication baked in from day one

---

🏗️ Architecture

graph TD
    subgraph Frontend ["Frontend (Next.js 14 + React 18)"]
        W["Chat Widget\n(Embeddable Component)"]
        UI["Chat UI\n(Ant Design + Tailwind)"]
        W --> UI
    end

    subgraph Backend ["Backend (Python / FastAPI)"]
        API["REST API\n/api/v1/chat"]
        AUTH["JWT Auth\nMiddleware"]
        HIST["Conversation\nHistory Service"]
        EMB["Embedding\nService (bge-m3)"]
        API --> AUTH
        API --> HIST
        API --> EMB
    end

    subgraph VectorDB ["Vector Store"]
        CHROMA["ChromaDB"]
    end

    subgraph DB ["Database"]
        SQLITE["SQLite\n(dev)"]
        POSTGRES["PostgreSQL\n(prod)"]
    end

    subgraph LLMs ["LLM Providers"]
        OAI["OpenAI\nGPT-4o"]
        GEM["Google\nGemini"]
        ANT["Anthropic\nClaude"]
        DS["DeepSeek"]
    end

    UI -->|"SSE Stream"| API
    HIST --> SQLITE
    HIST --> POSTGRES
    EMB --> CHROMA
    API -->|"Adapter"| OAI
    API -->|"Adapter"| GEM
    API -->|"Adapter"| ANT
    API -->|"Adapter"| DS

Loading

---

🚀 Quick Start

Prerequisites

• Python 3.11+
• Node.js 20+
• uv (Python package manager)

1. Clone the repository

git clone https://github.com/Raphasha27/ai-chatkit.git
cd ai-chatkit

2. Backend setup

cd backend

# Install dependencies via uv
uv sync

# Copy and configure environment variables
cp .env.example .env
# Edit .env — set LLM_PROVIDER and the matching API key

# Start the development server
uv run uvicorn app.main:app --reload --port 8000

The API will be available at http://localhost:8000.

3. Frontend setup

cd frontend

# Install dependencies
npm install

# Copy and configure environment variables
cp .env.example .env.local
# Edit .env.local — set NEXT_PUBLIC_API_URL=http://localhost:8000

# Start the development server
npm run dev

The chat UI will be available at http://localhost:3000.

---

⚙️ Environment Variables

Configure the root .env (or backend/.env) using the table below. Copy .env.example as your starting point — never commit real secrets.

Variable	Required	Default	Description
PORT	✅	8000	Port the backend API listens on
LLM_PROVIDER	✅	openai	Active LLM provider: openai | gemini | anthropic | deepseek | dashscope
OPENAI_API_KEY	✴️	—	OpenAI API key (required when LLM_PROVIDER=openai)
GEMINI_API_KEY	✴️	—	Google Gemini API key (required when LLM_PROVIDER=gemini)
ANTHROPIC_API_KEY	✴️	—	Anthropic API key (required when LLM_PROVIDER=anthropic)
JWT_SECRET	✅	—	Secret used to sign JWT tokens — generate with openssl rand -hex 32
DATABASE_URL	✅	sqlite+aiosqlite:///resource/database.db	SQLAlchemy async database connection string
EMBEDDING_MODEL	❌	bge-m3	Ollama embedding model for RAG features
CHROMA_PATH	❌	resource/chroma_db	Relative path for ChromaDB persistence
DEBUG	❌	false	Enable verbose debug logging

✴️ Only the key matching the active LLM_PROVIDER is required.

---

🗺️ Roadmap

• Unified provider adapter layer (OpenAI, Gemini, Anthropic, DeepSeek, Qwen)
• SSE streaming responses
• Persistent conversation history (SQLite + PostgreSQL)
• JWT authentication middleware
• ChromaDB + bge-m3 embedding integration
• Next.js 14 + Ant Design chat UI
• Built-in RAG pipeline with document upload
• Embeddable <ChatWidget /> npm package
• OpenAPI / Swagger documentation endpoint
• WebSocket transport option (alongside SSE)
• Admin dashboard for conversation analytics
• One-click Vercel + Railway deployment templates
• MCP (Model Context Protocol) tool integration

---

🤝 Contributing

Contributions, issues, and feature requests are welcome!
See CONTRIBUTING.md for guidelines and CODE_OF_CONDUCT.md for community standards.

---

🔐 Security

Found a vulnerability? Please read SECURITY.md before disclosing publicly.

---

📄 License

This project is licensed under the MIT License — see LICENSE for details.

---

_{Built with ❤️ by Koketso Raphasha · © 2026 Kirov Dynamics Technology}