Suggestify

A self-hostable project that generates weekly movie recommendations and sends them via email using Cloudflare Workers.

Requirements

Bun (JS Runtime), A Cloudflare account, Google AI and Resend API keys.

Services Used

Cloudflare Workers, Google's Generative AI (Model: Gemini 1.5 Pro), Resend (to send emails), and Cloudflare KV (to store previous responses and improve the new ones).

Setup & Config

Clone the repository: git clone https://github.com/zxcodes/Suggestify.git
Install dependencies: bun install
Create a .dev.vars file in the root and store your API secrets there. An .env file won't work here since workers don't run in a typical Node environment.

RESEND_API_KEY: Your Resend API key
API_KEY: Your Google AI API key

Create a KV namespace called SUGGESTIONS_STORE using Wrangler or CF Dashboard.
Configure your wrangler.toml file with the necessary bindings and environment variables (Refer CF Docs).

Development

Start the dev server by running bun dev and hit the cron by curl http://localhost:8787/__scheduled

Deployment

Authenticate with Cloudflare: wrangler login
Deploy the worker: bun deploy
To View logs from the deployed worker, do bun logs
The .dev.vars file is only for local dev. Use wrangler to upload your secrets to CF so that they're available when you deploy your worker. You can add secrets by running wrangler secret put <KEY>