Why I built this
A few months ago, I had zero coding experience. I got frustrated that every AI feed tool stores your personal taste data on their servers — data that essentially becomes their asset, not yours.
So I decided to learn to code and build something different.
What I built
My AI Curator — a Bluesky feed viewer where ALL your preference data stays in your own browser. Nothing goes to a server.
👉 Live: my-curator.vercel.app
💻 Code: github.com/etehero90-web/my-curator
How it works
- You select topics you care about (Technology, Design, Gaming, etc.)
- You can add custom keywords (k-drama, climate, yoga — anything)
- The app filters your Bluesky feed and shows matching posts first
- All your preference data is stored in your browser's localStorage only
Privacy features I'm proud of
Pod access On/Off switch — a master toggle that instantly blocks all data reading. When it's off, the app shows posts chronologically with zero personalization.
Full access log — every single time the app reads or writes your data, it's logged with a timestamp. You can see exactly what happened and when.
Granular controls — toggle read history and topic preferences independently.
JSON export/import — download all your preference data as a JSON file anytime. Take it anywhere.
The tech stack
- Pure HTML/CSS/JavaScript — single file, no framework
- Bluesky public API (no login required for public feeds)
- Browser localStorage for all state management
- Zero backend, zero database, zero cookies
What I learned
Starting from zero, the biggest lesson was: just start. The first line of HTML felt impossible. Now I have a working app that real people can use.
The second lesson: build for a real problem. I genuinely wanted this tool to exist, which kept me going when things got hard.
What's next
- Support for multiple Bluesky accounts simultaneously
- Mastodon/Fediverse support
- Better onboarding experience
Would love any feedback — especially on the code quality since I'm still learning!
Top comments (0)