AskIt

AskIt is a premium, high-performance AI web application featuring an advanced multi-modal agentic architecture. It seamlessly integrates Retrieval-Augmented Generation (RAG), Model Context Protocol (MCP), and vision-based reasoning within a sleek, neuromorphic user interface.

Built entirely on the modern Next.js 15 App Router and powered by Bun, AskIt utilizes Supabase as a robust vector store and authentication provider, creating a secure and highly scalable conversational AI platform.

Features

• Multi-Modal Agentic AI: Supports complex, multi-step reasoning using tools (like search_documents) and natively processes both text and image inputs (vision API).
• High-Performance RAG: Automatically chunks, embeds (via OpenAI), and stores uploaded documents in a Supabase pgvector database for low-latency semantic retrieval during chat.
• Neuromorphic Design System: A meticulously crafted visual identity featuring a black, pink, and white palette, smooth Framer Motion animations, and beautiful typography (Playfair Display and Inter).
• Secure Authentication: Integrated Supabase Auth offering both traditional email/password and seamless Google OAuth sign-in.
• Production-Ready Dockerization: Highly optimized standalone output configuration for easy containerization and deployment to modern cloud platforms like AWS App Runner or ECS Fargate.

Architecture & Stack

The application stack is chosen for maximum developer velocity and production performance:

• Runtime & Package Manager: Bun
• Framework: Next.js 15 (React 19, Turbopack)
• Database & Auth: Supabase (PostgreSQL, Auth, Row Level Security, pgvector)
• AI & Orchestration: Vercel AI SDK, OpenAI (gpt-4o, text-embedding-3-small), Groq, Google Generative AI
• Styling & UI: Tailwind CSS, Framer Motion, Neuromorphic design principles

Repository Structure

The repository is organized to separate the public landing experience from the secure application core.

• src/app/ — Next.js App Router routes including the landing page (/), authentication pages (/login, /signup), and the core application interface (/app).
• src/app/api/ — API routes handling the streaming chat interface, multimodal processing, and document ingestion (chunking, embedding, storing).
• src/lib/ — Core utilities including the Supabase client logic and the RAG engine handling pgvector matching.
• supabase/migrations/ — Essential SQL schemas defining profiles, documents, chunks, chats, messages, Row Level Security (RLS) policies, and the match_chunks RPC function.

Getting Started

1. Installation

Ensure you have Bun installed, then clone the repository and install dependencies:

cd askit
bun install

2. Supabase Configuration

1. Create a new project at supabase.com.
2. Navigate to the SQL Editor and run the initial migrations sequentially:
   • supabase/migrations/00001_initial.sql
   • supabase/migrations/00002_match_chunks.sql
3. Enable Google under Authentication → Providers and add your Google OAuth credentials.
4. Set the redirect URI to https://<your-project-ref>.supabase.co/auth/v1/callback.
5. Retrieve your Project URL and anon (public) key from Settings → API.

3. Environment Variables

Create a .env.local file in the root directory:

NEXT_PUBLIC_SUPABASE_URL=https://your-project.supabase.co
NEXT_PUBLIC_SUPABASE_ANON_KEY=your-anon-key
OPENAI_API_KEY=sk-...
# Add other required API keys (Groq, Google, etc.) as needed.

4. Running Locally

Start the development server using Turbopack for ultra-fast compilation:

bun run dev

The application will be accessible at http://localhost:3000.

Deployment

AskIt is fully configured for Dockerized deployment, making it trivial to host on AWS.

Building and Running with Docker

docker build -t askit-app .
docker-compose up

Deploying to AWS App Runner

AWS App Runner provides a zero-management container deployment experience.

1. Create a repository in the AWS ECR Console named askit.
2. Authenticate your local Docker daemon to your AWS ECR registry.
3. Tag and push the askit-app:latest image to your newly created ECR repository.
4. Navigate to AWS App Runner, select Container Registry -> ECR, and point it to the askit:latest image.
5. Set the port to 3000 and map your runtime environment variables (from .env.local).
6. Click Create Service.

Important: Once deployed, update your Supabase Site URL and Google Cloud Authorized JavaScript origins with the generated App Runner public URL.

License

MIT License