Feedle: A search engine for blogs and podcasts | Hacker News Reader

Feedle: A search engine for blogs and podcasts

feedle.world

292 pointsmicroflash68 comments

The world needs blogs, repositories of long-form original thinking on the Internet. As a blogger I appreciate the opportunity to place my link. I salute your initiative and wish you success!

True,

And it will be great when a service such as this has a auto-crawler, categorizer LLM that takes all the content in the repo and groups them together based on content context tagging - so a diasproa of blog posts from anyone/where about [TOPIC] could be groupded, and even compared...

"Show me all the blog posts on [TOPIC] but show me the top opposing opinions on said [TOPIC]"

Or "show me all the folks who have consensus on [SUBJECT]"

etc.

That sounds very useful. I wonder if that might be more in the wheel house of a search engine like the folks at Kagi? I'm trying to think of what is the closest thing to this that exists.

Gather what properties you can dissect, sort the props of the search results by how many share them, display the most frequent x with a more button.

The “Submit” link just takes me to the Tally home page. I wonder if the link is broken of if they exceeded a limit.

This is a known bug. There are form links elsewhere in these comments.

I agree with you.. but there is a lot of issues related to blogging that has caused it's decline in popularity. (I.e. content aggrigation, cross connection social comment features, reduction of spam in discovering/blog reputation, etc)

For me I have a blog but I'm struggling to write for it because I don't have an audience that makes the blog valuable for me.

Would love to hear if you think there's a solution to this.

I have some ideas, which I wrote about here: https://github.com/yakkomajuri/recess/blob/main/manifesto-is...

Curious if you think this makes sense.

i've had dozens but just write for yourself. I find it hilarious and humbling but the most enjoyable is ironing out the inconsistencies or gaps in my own thoughts. Have the cake or eat it? It can be left as an open question too. Asking the right questions is useful.

Every year or so someone will reach out to me based on something I wrote. It’s really gratifying to solve someone’s problem.

I wrote a little blog program because I find the blog format (articles sorted by date) convenient for reading about things someone did, so I figured when I do things I should also put them in that format. But it turns out I don't do a lot.

I have written a dozen blog engines in whatever language caught my fancy at the time. In the end I realized I'm writing my blog for myself, not for an audience. The value is in sorting my mind, not someone reading what I wrote. I have moved off online publishing since then. Now I just journal in Zim desktop wiki or even on paper and you and I are better off for it. I'm just guessing, but maybe question your desire to have an audience.

Similarly my band refuses to release an album until we have a huge audience!

Seriously though, write for you! The eyeballs will come later

You call them issues but I prefer to think of them as a filter that keeps blog content high value exactly because it discourages those who are not primarily interested in sharing information but are after some other goal.

A bit tangential, I understand that Google is not well-received, and rightly so in some regards. However, has anyone been using something like this with Google: "<search-terms> site:blogpost.com OR site:github.io OR site:wordpress.com OR site:medium.com OR site:substack.com"?

And occasionally throwing in reddit.com and news.ycombinator.com as well. Google seems to be slightly biased towards github.io and medium.com, but I've quite recently found that it is a great way to browse the web.

The above search command is similar to what these niche search engines are doing but gets much more diverse results. Sure, it is likely to miss certain niche blogs. However, this is where adding site:news.ycombinator.com comes to the rescue - good niche blog posts are highly likely to eventually appear here.

It is far from a panacea, but I have found the experience to be quite decent. I have this command lying around in my shell and notes, so I just copy-paste it and add on to it.

> However, has anyone been using something like this with Google: "<search-terms> site:blogpost.com OR site:github.io OR site:wordpress.com OR site:medium.com OR site:substack.com"?

Besides the privacy aspect of the service, this is almost the whole value prop of Kagi: you can mark certain sites as ranking higher, lower, or banishing them from your results entirely. I never thought I'd stop using Google but I've become a happy Kagi customer.

Why would you explicitly ask google to only include the low value blog hosts?

Since the creators are here: how are you going to defend against SEO spam? Curation when adding the blog to the index wont work as anyone could make a reasonable-looking blog with AI generated articles in just a few hours and after you've added it to the index push a ton of aggressive spam.

What kind of queries do you handle well? Do you index podcast transcripts?

I adore the idea of a podcast/blog search engine! And results can be streamed via RSS, wonderful! I use Feedly to subscribe to important RSS feeds, including specific Hacker News authors.

Current database is a bit light: it didn't have "Science Vs", a popular podcast. Added.

For comparison, https://www.podchaser.com/ only has podcasts but it's more fleshed out.

I highly recommend https://podcastindex.org/ It's open and can be used by many. It has all the podcast I need. It also powers a lot of open source podcast apps like https://antennapod.org/

Hello everyone, co-creator of Feedle here. Thank you so much for letting the world know about our tiny side project! While I tred answering some of people's pressing questions on this thread, I think they deserved a more thorough explanation, so I wrote a blog post instead: https://preslav.me/2024/03/25/feedle-on-the-hacker-news-fron...

Beweare, it's a bit lengthy, but hopefully, it would address some of your concerns and feedback. You can always reach out to me and the team with ideas and suggestions. Cheers!

Cool idea, I like that you can follow the queries as feed.

The crux will be maintaining a well-curated/reviewed list of feeds.

Also: the submit form seems to be broken - leads to the landing page of some form builder saas.

Minor feedback: After doing a search with 0 results, there is a "load more" button that doesn't do anything.

Awesome project, thanks for putting it together and sharing it with us!

A suggestion: allow users to flag dead links for removal.

I searched for “task management”. One or more results led to pages that are 404/no longer exist. I’d like to help you remove those results. My suggestion seems appropriate because users add Feedle’s links in the first place.

internet archive links

Thank you, although I know how to access the links through an archive service if I want to. My point was different: dead links are cluttering up the search results (and therefore the RSS). It would be nice to flag them for removal (or if the service removed dead links programmatically).

As an aside for the developers, I find that the searches contain a lot of noise. (For instance, a search for OmniFocus turns up results from several sites that only contain summaries of release notes for minor versions of the app.)

Could be nice to have an “I’m feeling lucky”

stumbleupon 2.0

Same as high waisted wide jeans, everything is coming back

It's very high on our list. Optimistically, we are going to release that feature by the end of next week.

https://indieblog.page/random

Something similar

This is surprisingly useful. I was trying to find a very specific podcast about a type of data algorithm and google podcast search just recommended me the wrong type of podcast or couldn't find any. Tried this and it found 9 podcasts that matched what I was after.

Edit: OK they were all articles and not podcasts, so it's possible no one has a podcast about it, but this is still useful

I'd like to submit my blog, but I can't.

The "Submit your blog or podcast" button at the top of the page takes me to an advertisement for an unrelated website.

I found the answer in a different reply:

https://news.ycombinator.com/item?id=39801350

The link to submit a new blog or podcast doesn't work, it redirects to the homepage of a form builder service.

It's a known issue for a portion of our visitors, which we are working on. For the time being, you can use the direct link to the form: https://tally.so/r/mJ11E7

This is in some ways similar to something I built recently:

https://app.recessfeed.com/

We seem to have similar ideas, so would love to have a chat! This line aligns a lot with what I have in mind:

"Discover content creators you may not know about"

As with the majority of the recent search startups: no hiragana, no hebrew, no cyrillic.

Co-creator here: we have multi-language support on the backend site, but it is still in a bit of an experimental stage at the moment. Not all posts are being attributed the right language, so we are still experimenting there. For the time being, EN-language blogs are our first priority. That said, we will beta-launch the multi-language support very soon.

A search for "Kubernetes" shows a ton of corporate "blogs".

Thanks so much for building this. Very cool. In addition, making clear that the feeds are embeddable in Obsidian is just great. Will see how well this searches and will start using today!

Found some interesting blog posts using this, thanks!

I dont know if you want to monetize this. But I suggest make it open source.

You will receive tons of improvements and updates. I love this idea!

To the creators: I've already read three very interesting blog posts that I never would have read otherwise. Well done.

Cool idea, does a blog need to be submitted? There's a movie blog I didn't see in the search

There's a submit button in the header. Top right on desktop.

That's quite cool. How are feeds added? Is it all manual curation and submissions?

There's a link in the header to submit your own feed. Sadly that link didn't work for me.

Seems fixed now, I just submitted a podcast feed.

It's a known issue for a portion of our visitors, which we are working on. For the time being, you can use the direct link to the form: https://tally.so/r/mJ11E7

Is there a search engine for search engines yet?

What are you using to index the blogs and podcasts?

And do you permit people to submit blogs and podcasts that they don't own? That's unclear at the moment.

damn shame there doesn't seem to be a search facet for saying "only podcasts" since I don't need another search engine for webpages

I must confess the results don't seem too relevant for my first search: as I recently lost a family member to it, I searched "philosophy suicide", and got the following:

* Post-Human Capitalism and Revolution: Detroit and Blade Runner 2049

* The voice in your head: How a movement of people who hear voices is reshaping our understanding of mental illness – and consciousness itself.

* The Ghosts of Mark Fisher

* The radicalism of Randolph Bourne: Bourne’s affinity with outsiders drove his vision of making North America a united states of communities. A century on, his writings have become more relevant than ever

* Is the quest for immortality worse than death?

* The Dangers of Meritocracy

So far only the 4th result seems more than just tangentially relevant, and I know of much more relevant mainstream podcast episodes specifically dealing with the philosophy of suicide, but none of those seem to come up.

It's a cool idea but I really don't like the front-end being used in the site. For instance, if you try to select the footer element, to block it, instead of the site showing you "#footer" (or something legible), it shows you something like ".lg\:px-8.sm\:px-6.px-4.max-w-7xl.mx-auto > .py-16". As well as it adding a "?ref=feedle.world" in all links you click by default. Also it would be nice an option that automatically loads more when you scroll down.

Oh, and that "Welcome to feedle!" popup seems like dark pattern to me.

I would have said the same thing about infinite scrolling; that always feels like a dark pattern to me because it makes it hard to find specific results again, hard (or impossible) to get to any links in the page footer, etc.

I always just want a single page that I can link back to later :)

To me, I think it depends on what is the goal when you are searching something and what is that data in particular. Infinite scroll and pagination offer different user experience proposes, and make different assumptions about the user.

To give a blatant example where I don't see a pagination system making that much sense, pagination on a personalized news feed of some social media. That information is changing constantly. It's a never-ending information stream.

Even in the case here, if you are searching for a blog or podcast, you could just link that blog article rather than the page search where you found that article. Especially in a situation where you don't have any interaction with that content aside clicking on it. Especially when the search engine doesn't seem to have millions of results, which seems to be the case here, it seems to be a relatively small list of blogs submitted to this site.

Infinite scroll on the other hand have some problems especially in situations where you are expect to interact with that content one way or the other (such as a commentary section). In fact, when dealing with really large amounts of data, it essentially buries that data. Think about a commentary section with 60.000 commentaries. Essentially 98% of them are hidden way because the browser will crash before loading them. So if you decide to go down infinite scroll route, I think the very least you could do would be offer another way to search for stuff, such as search option inside the commentary section (although nobody does that). But I think a pagination system would be better in that situation.

Anyway, just some food for thoughts.

IMO pagination makes sense with feeds when a feed can be sorted in some ordering. I'd rather jump to page 6 instead of scrolling and scrolling and scrolling to get back.

Actually because of the infinite scroll and basically mental way that social media works I gave it up...

No, it's not constantly changing. There was a publication at rhw certain time and it can be shown at that time (ideally you can have inbox with unread entries). Right no it's impossible to follow anything in any sane manner...

to that end I use only RSS and from social perspective I added a couple of reddit sub feeds there and that's all...

This is the end result of the tailwindification of the internet.

Its TailwindCSS. Classes are CSS, can’t really match those.

What's the use case, in your opinion, for an average user to inspect the footer? Or any element on that website for that purpose?

Average users can't or don't need to be helped. Let's think more about those who have special needs.

Is the assumption here that Tailwind classes are somehow not A11Y compliant? That's false, you can both use Tailwind and make website fully accessible, reader-compatible.

When you try to alter the style of a calendar, in order to be able to read the text more clearly, if making the style of the date readable turns the title into garbage (because there's no semantic information, only they happen to share the same visual class), then someone is going to have a hard time using your website because of your decisions not taking into account their needs.

(This is a slightly tweaked real life example; it was adjusting the style for printing so that the calendar was accessible offline.)

The icon color and name made me think this was Feedly not Feedle, took a moment to understand they were unrelated

Reply on news.ycombinator.com