Exam Preparation Agent Workshop

What's Inside

• FastAPI service that is backed by OpenAI Agents.
• ChatKit Web Component wrapped in React with a document panel.
• Vector-store tooling for ingesting documents and exposing REST endpoints for previews, uploads.

Prerequisites

• Python 3.11+
• Node.js 22+ (Use nvm)
• uv
• OpenAI API key as OPENAI_API_KEY in .env

Steps to Run:

Setup Notion

1. https://github.com/makenotion/notion-mcp-server/blob/main/README.md#1-setting-up-integration-in-notion
2. https://github.com/makenotion/notion-mcp-server/blob/main/README.md#2-connecting-content-to-integration
3. Copy Internal Integration Secret and put in your .env

Setup Logfire

1. https://logfire-us.pydantic.dev/login
2. Signin using preferred method
3. Click "Lets go" to use default starter-project
4. Copy Write token and put in your .env

Start the FastAPI backend

1. Setup environment
   • Copy the template environment file into your own .env file: cp .env.template .env
2. Create or reuse a vector store:
   • Visit OpenAI Vector Stores and create a vector store, or, Use existing vector store
   • Copy the Vector Store ID (e.g. vs_abc123) and set in your .env file, alongside your OpenAI API Key
3. Use nvm OR ensure Node V22+: nvm use
4. Install dependencies and launch the API: npm run backend

Start the React Frontend

1. Use nvm OR ensure Node V22+: nvm use
2. Launch frontend server: npm run frontend