Let Me Ask API

An API developed during NLW Agents by Rocketseat for live streams and broadcasts, which uses an AI agent to predict answers to questions in real time.

🧪 This is the first version of the project, exactly as taught during the event. I intend to implement improvements in the future or use it as a base for similar projects.

📑 Table of Contents

• Technologies
• Setup and Configuration
• Running the Project
• API Endpoints
• Front-end
• Author

Technologies

The project was developed using the following stack:

• Node.js + native TypeScript (experimental strip types) — modern and typed development
• Fastify — lightweight, fast, and efficient web framework
• PostgreSQL + pgvector — data persistence with support for vectors and vector search
• Drizzle ORM — type-safe operations on the database
• Zod — robust schema and data validation
• Docker — database containerization for easy setup and deployment
• Biome — code linting and formatting to maintain quality and consistency

Setup and Configuration

Below is a step-by-step guide for running this project.

1. Prerequisites

To run this project, make sure you have the following installed and configured on your system:

• Node.js (use a version that supports --experimental-strip-types)
• Docker and Docker Compose

2. Clone this Repository

Example using SSH:

git clone git@github.com:MarcosAntonio15243/let-me-ask-api.git

3. Install Dependencies

For this project and documentation tutorial, we will use the NPM package manager. To install the dependencies, run the following command at the root of the cloned repository:

npm install

4. Docker and Database

To set up the PostgreSQL database with the pgvector extension, run the following command at the root of the repository:

docker-compose up -d

💡 Note: Make sure no application is using port 5432, as it is the default database port for this project.

• If the port is already in use, you can temporarily stop the service using it, or
• Change the port in the docker-compose.yaml file to one that is available on your machine.

5. Configure Environment Variables

Create a .env file at the root of the project with the following format:

# HTTP API port
PORT=
 
# Database URL
DATABASE_URL=
 
# Gemini API key
GEMINI_API_KEY=

6. Run Database Migrations

To run the database migrations, execute:

npx drizzle-kit migrate

7. (Optional) Seed the Database

If you want to seed the database with initial data, run:

npm run db:seed

Running the Project

Once all setup steps are completed, run the following command to start the project:

• For development environment:

  npm run dev

• For production environment:

  npm start

API Endpoints

You can explore and test the available API endpoints using the documentation below.

The default base url is: http://localhost:3333.

• GET /health - Verifies if the API is running properly.
• GET /rooms - Retrieves a list of all available rooms.
• POST /rooms - Creates a new room.
• GET /rooms/{roomId}/questions - Retrieves all questions from a specific room.
• POST /rooms/{roomId}/questions - Submits a new question to a specific room.

Front-end

Want to try the interface?

Check out the web app here: Let Me Ask Web

Author

Made by Marcos Antonio.

• 💻 Full Stack developer dedicated to building complete solutions by combining modern, functional user interfaces with robust back-end architectures.
• 📚 This project represents my first experience using tools like Drizzle ORM and the pgvector extension, with a focus on AI-powered APIs based on multimedia content.
• 🚀 Always open to feedback, collaboration, or ideas for improvement!
• 📫 Feel free to connect with me on LinkedIn or check out more of my projects here on GitHub.