Hacker News

Show HN: Bookmarklet to save NotebookLM audio into a personal podcast

Hacker News - Tue, 03/11/2025 - 6:30am

I've been listening to lots of audio overviews on Google's NotebookLM lately (having a baby strapped to you forces new media consumption habits..!) & got annoyed they weren't in my podcast app with all my other podcast-y content.

So I made a bookmarklet which works to upload the audio then converts it to MP3 and adds it to a private personal RSS feed which you can add to any podcast app with a click or by pasting the URL.

The bookmarklet's javascript obviously, the backend uses Python in Cloud Run for conversion then Laravel with Livewire for the UX, hosted on GCP, and Paddle for subs (but is free to try)

Would love any feedback.

Comments URL: https://news.ycombinator.com/item?id=43331040

Points: 1

# Comments: 0

Categories: Hacker News

Show HN: Instantly Extract Color Palettes from Any Website

Hacker News - Tue, 03/11/2025 - 6:30am

Hey HN!

I built https://colorize.design, a free tool that extracts color palettes from any website in seconds. Just enter a URL, and it analyzes the site's visuals and code to generate curated color palettes.

Why I Made This

As a designer, I often come across beautifully designed websites and want to quickly grab their color schemes. Inspecting elements or using browser extensions felt tedious, so I built a tool to automate the process.

How It Works

Screenshot Analysis: Extracts dominant colors from the site’s homepage.

Favicon Extraction: Pulls key colors from the site's logo.

Code Parsing: Scans HTML, CSS, and JS files to find color declarations.

Smart Grouping: Uses a nearest-color algorithm to remove duplicates and improve usability.

Editing & Exporting: Lets users refine palettes and download them in multiple formats (CSS, SCSS, design tools, etc.).

It’s powered by Puppeteer for web scraping, ColorThief for image analysis, and custom regex to extract code-based colors. The biggest challenge was handling messy websites—redirects, popups, cookie banners, and huge stylesheets all made this trickier than expected!

What’s Next

I’m working on: A Chrome extension for one-click extraction. Enhanced API features for developers.

Would love to hear your thoughts! What would make this more useful for you? Let me know what you think.

Comments URL: https://news.ycombinator.com/item?id=43331038

Points: 2

# Comments: 1

Categories: Hacker News

Intel Setting the Stage

Hacker News - Tue, 03/11/2025 - 6:22am
Categories: Hacker News

To designers designing hardware products

Hacker News - Tue, 03/11/2025 - 6:18am

Which part takes the most time when designing a hardware product?

Comments URL: https://news.ycombinator.com/item?id=43330976

Points: 1

# Comments: 0

Categories: Hacker News

Any guide to migrate from ssojet to auth0

Hacker News - Tue, 03/11/2025 - 6:14am

I am using ssojet currently and facing some problems to connect my microsoft active directory, they support azure active directory but they do not have out of the box support for active windows active directory.

I am look for any helpfull guide to migrate from ssojet to auth0, I can’t afford enterprise plan.

Comments URL: https://news.ycombinator.com/item?id=43330948

Points: 1

# Comments: 0

Categories: Hacker News

Ask HN: How Is AI SEO (Geo) Different from SEO?

Hacker News - Tue, 03/11/2025 - 6:12am

We've recently started to think more seriously about SEO as a growth channel.

Since we're already getting multiple leads that self-attribute to AI tools like Perplexity and ChatGPT, we want to optimize for more of this. There are now tools that offer to help you do SEO for AI (GEO, generative engine optimization).

But what I'm wondering about is this:

Both ChatGPT and Perplexity ultimately use Google's or Bing's search index, which makes me think optimizing for conventional search engines is the way to go because it achieves the same thing?

The two differences I can imagine are this:

a) For non-search AI chats, getting included in the training data could be another objective, though the corpus of training data will ultimately surface the incumbents in any space.

b) An AI search product might use different queries that humans don't actually search for much (so there's little search volume), so optimizing for them could be more useful than before.

Have any of you made other experiences or know more?

Comments URL: https://news.ycombinator.com/item?id=43330935

Points: 1

# Comments: 0

Categories: Hacker News

The Component Gallery

Hacker News - Tue, 03/11/2025 - 6:07am

Article URL: https://component.gallery/

Comments URL: https://news.ycombinator.com/item?id=43330901

Points: 1

# Comments: 0

Categories: Hacker News

Pages