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

1. Clone the repository: git clone https://github.com/zxcodes/Suggestify.git

2. Install dependencies: bun install

3. 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

4. Create a KV namespace called SUGGESTIONS_STORE using Wrangler or CF Dashboard.
5. 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

1. Authenticate with Cloudflare: wrangler login
2. Deploy the worker: bun deploy
3. To View logs from the deployed worker, do bun logs
4. 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>