AI Getting Started

Free

A Javascript AI getting started stack for weekend projects, including image/text models, vector stores, auth, and deployment configs ![GitHub Repo stars](https://img.shields.io/github/stars/a16z-infra/ai-getting-started?style=social)

FreeFree tier

Inputs: text, imageOutputs: text, image

Type

Open Source

Company

a16z-infra

Visit Website GitHub (40 stars)

About AI Getting Started

AI Getting Started is an open-source JavaScript AI stack from a16z-infra designed to accelerate weekend projects. It integrates Next.js for app logic, Clerk for authentication, Pinecone or Supabase pgvector for vector storage, Langchain.js for LLM orchestration, Replicate for image generation, OpenAI for text models, the AI SDK for streaming, Fly for deployment, and Arcjet for security. The repository provides a pre-configured template with example blog data for Q&A, environment setup scripts, and Docker support, enabling developers to quickly build and deploy AI applications with image/text models, vector search, and authentication.

Key Features

Next.js application framework for frontend and backend

Clerk authentication integration

Pinecone or Supabase pgvector vector database support

Langchain.js LLM orchestration

Replicate image generation model

OpenAI text model integration

AI SDK for streaming responses

Fly deployment configuration

Arcjet security integration

Docker and fly.toml for containerized deployment

Pros & Cons

Pros

Comprehensive pre-integrated stack saves setup time
Open source and free to use
Includes both text and image model support
Supports multiple vector store options (Pinecone, Supabase pgvector)
Comes with example data and quickstart instructions
Deployment configs for Fly and Docker
Active community via Discord

Cons

Requires signing up for multiple external services (Clerk, OpenAI, Replicate, Pinecone/Supabase, Arcjet)
Learning curve for developers unfamiliar with the individual components
Not a production-ready solution out of the box; requires customization
Relies on proprietary third-party APIs that may incur costs beyond the free tier

Best For

Rapid prototyping of AI-powered applicationsBuilding chatbots with retrieval-augmented generation (RAG)Creating image generation web appsDeveloping AI-assisted Q&A systems over custom contentLearning and experimenting with modern AI stack components

FAQ

What services are required to run AI Getting Started?

You need API keys for Clerk (auth), OpenAI (text model), Replicate (image model), either Pinecone or Supabase (vector store), and Arcjet (security). All provide free tiers.

Can I switch between Pinecone and Supabase as vector store?

Yes, the template supports both Pinecone and Supabase pgvector. You can choose which one to configure by filling the corresponding API keys.

Is there a live demo?

Yes, the repository includes a link to a live demo deployed on fly.io.

How do I get started quickly?

Fork the repo, clone it, install dependencies, copy .env.local.example to .env.local, fill in the required API keys, and run the embedding generation script followed by the dev server.