Spezi AI Assistant

A friendly code assistant that helps users with their questions about the Stanford Spezi framework.

Features

• Next.js
• AI SDK
• shadcn/ui
• LangChain
• Data Persistence with Vercel Postgres and Vercel Blob
• NextAuth.js for authentication

Local Development

You will need to update the environment variables see the example file in .env.example to run the Spezi AI Assistant.

Note: You should not commit your .env file or it will expose secrets that will allow others to control access to your various OpenAI and authentication provider accounts.

Database Setup

1. Set up your Postgres database (either locally or using Vercel Postgres)

2. Add your database URL to the .env file:

   POSTGRES_URL="postgres://..."
   

3. Initialize the database:

   # Generate the database schema
   pnpm db:generate
   
   # Push the schema to your database
   pnpm db:push
   
   # Optional: Open the database UI
   pnpm db:studio

Ingesting Documentation

The system can automatically generate documentation from GitHub repositories and ingest it into the database:

1. Add the repositories you want to ingest to the config/repositories.yml file.

2. Generate LLM-friendly markdown summaries of the repositories that are saved into the resources directory:

   pnpm generate

3. Ingest all markdown files from the resources directory into the database:

   pnpm ingest

4. You can also generate and ingest in one command, if desired:

   pnpm generate-and-ingest

You can customize all of the above in scripts/ingest.ts.

Install dependencies

pnpm install

Run the development server

pnpm run dev