Build your own read-later app with Crawl4AI and Supabase

Using the Opensrcs tools, look up Crawl4AI and Supabase. In simple words, tell me what each one does and what I need to do to follow their licences if I build a small app with them.

Set up Crawl4AI on my computer using Docker (the official unclecode/crawl4ai image, on port 11235). Run the commands for me, check that it's working, and tell me in simple words when it's ready. I'm a beginner, so explain anything I need to do myself.

Build me a simple "read later" web app. I'm a beginner, so keep it simple, explain each step in plain words, and tell me exactly how to open it at the end.

What it does:
- A home page with a box to paste a web address and a "Save" button. Below it, my reading list, newest first, showing each article's title, website and the date I saved it.
- When I press Save, the app gets the article with Crawl4AI and saves it in Supabase. If it fails, show a clear message.
- Clicking an article opens a reading page with the article nicely formatted, a link to the original page and a Delete button.

How to build it:
- Use Next.js with TypeScript and plain CSS. Make it clean, easy to read and work well on phones.
- Crawl4AI is running on my computer at http://localhost:11235. To get an article, send a POST request to http://localhost:11235/md with the JSON body {"url": "<the address>", "f": "fit"}. The reply is JSON with the article text in its "markdown" field. Test this once with https://example.com first and adjust if the reply looks different.
- Store articles in a Supabase table called "articles" with: id, url (unique), title, markdown, created_at. Give me the SQL to create it with row level security turned on, tell me where to paste it in Supabase, and wait for me to confirm.
- Only talk to Supabase from the server, never from the browser. Use a .env.local file for SUPABASE_URL and SUPABASE_SECRET_KEY, tell me where to paste my keys, and make sure the file is never shared (add it to .gitignore).
- Use the article's first heading as its title, or the website's name if there isn't one.
- If I save the same address twice, update it instead of showing an error.
- Add a small "Credits" line in the footer with exactly this text: This product includes software developed by UncleCode (https://x.com/unclecode) as part of the Crawl4AI project (https://github.com/unclecode/crawl4ai).

If you have the Opensrcs tools, check the licences of Crawl4AI and Supabase first and tell me if I need to do anything else.

The app isn't working. Here's what I did and what I saw:

[describe what happened, and paste any error message here]

Please find the problem and fix it, and tell me in one or two simple sentences what was wrong.

Add a search box above my reading list so I can find articles by words in the title or text.

Add a "Read" checkbox to each article so I can mark what I've finished, and a way to show only unread articles. If Supabase needs a change, tell me exactly what to paste and where.

Add sign-in with an email link using Supabase, so each person only sees their own reading list. Keep everyone's articles private from each other, and tell me exactly what I need to paste into Supabase.

Using the Opensrcs tools, suggest two or three open-source projects that could add a fun new feature to my read-later app, like summaries or reading articles out loud. For each one, tell me what it adds, its licence, and whether it costs anything.