Hacker News new | past | comments | ask | show | jobs | submit login
Winds – An Open Source Personalized RSS Reader (getstream.io)
161 points by tortilla on Nov 11, 2016 | hide | past | favorite | 111 comments



Do people really want recommendations?

I personally want to just have an RSS reader that gives me the articles from the sites I freaking asked for. Not what some service thinks I'll like. Google, Pocket, and other sites are already doing a great job of thumbing through what I read and trying to sneak in their "suggestions": that's the kind of invasive data-combing that I want to avoid.


There are a few distinct use cases for RSS readers with very different requirements for automatic filtering.

I use RSS readers to keep up with infrequently-updated fiction and webcomics. In this use case, "suggestions" are worse than useless - if I miss a daily or weekly strip or, worse yet, a chapter being posted, that can cripple my enjoyment of the story. I check my RSS feeds once ever two or three days and read exactly every single post in them, all twenty or forty of them.

There are other people (myself about six years ago, for example) that use their RSS reader as a news firehose. They get a hundred posts an hour and even checking their titles manually would be difficult. In this case suggestions are not just a good idea but actually required.

Personally, I'd like an RSS reader that has a per-feed toggle for automatic filtering. I so far haven't seen that.


And even both at the same time. I have a set of comics and low-frequency posting blogs that I want to always read every post, in order, and defer them if I didn't want to read it right now. And then a firehose of other articles that I may or may not be interested in reading and can skim it to see if I want to read it, and then open it in a new tab and continue.

I feel like it's challenging to get both of these use cases in the same app. I could do it in google reader by having categories, but when it closed I switched to Inoreader and it does the "helpful" thing of only showing the last 30 days if you want to read it in chronological order, and used up my mental energy switching and basically just gave up RSS after that and haven't read them in a long time.


For webcomics, I just want to plug this site: http://piperka.net/

(Free and mostly open source: http://piperka.net/blog/2015/about-crawler-source/, and currently being rewritten from unmaintainable Perl into Haskell)


For macOS, there's Vienna:

http://www.vienna-rss.org/

I'm in a similar situation regarding use case - some busy news feeds, where I'll just scan the titles and read a few articles of interest, and plenty of comics that aren't updated daily, and thus can't slot conveniently into an OmniWeb workspace.


Newsblur has per-feed filtering configuration.


Winds has 2 displays. One is personalized, the other is chronological. Thanks for explaining how you use RSS.


https://github.com/swanson/stringer -- I built an RSS reader a few year agos with the same ethos: "Stringer has no external dependencies, no social recommendations/sharing, and no fancy machine learning algorithms. But it does have keyboard shortcuts..."

It's not being actively developed (it's 'feature complete' and working just fine), but might suit your needs.


Not being a dev that is a pain to install, would have been nice to try it.


Click the big purple "Deploy to Heroku" button. If you have an account, it will basically start running for you.


Happy user here :-) has been running smoothly for over 2 years already.


Thanks for posting, this looks great and I'll probably use it.


I don't care much about being suggested more feeds to follow. Instead, I wish the learning algorithm would notice when I don't bother reading certain things (granular below the whole-feed level) in my RSS, and move them out of the way.

I've been pondering setting up an RSS-to-email forwarding gateway, and a separate Gmail account, so that I could use Gmail's online-learning mail categorization system to clean up my feeds for me. If a service can do any better than that, I'd like to see it.


NewsBlur has a training feature which allows you to do things like that—excluding items with a certain tag, or a certain keyword in the title. How it works seems a little opaque, but it does seem to work.


We track both impressions and engagements so that allows us to see when you're not interested in something. Details about the integration: http://blog.getstream.io/personalization-machine-learning-ne...


I felt the same way until I had the following problem:

1. By subscribing to something that always had good content I missed quite a bit from occasional interesting posts on other feeds. 2. Subscribed to those occasionally good feeds. Now I have way too many feeds. 3. Screw it time to just use HN, social media and reddit. 4. Miss interesting articles not on those networks. 5. Now I use both rss, email lists, HN, social media/give up and miss interesting articles.

I would use a product that was a really good classifier for articles I would like and aggressively found them for me. I don't have enough time in the day to read through every blog post someone writes.

Cooler yet if the trained model were portable and I could use it in other applications.


I used to love Google Desktop Search for this feature:

It would create an automatically updated stream of articles based on (I guess) RSS feeds of recently visited sites.

Sometimes someone asks what we would pay for.

I think I would pay USD25 without much hesitation to get that thing back exactly as I remember it.


I feel that way about a lot of google products. If they combined that tech with reader (RIP old friend) I would pay monthly.


Zite used to do this for me. Hope we'll get Winds to that level.


That's my big takeaway from the last week. I don't want algorithms influencing how I think.


Yeah, that's why I don't use newsblur for instance.

I want to see EVERYTHING that are in the feeds I follow.


We do that with Nextcloud News: https://github.com/nextcloud/news/#nextcloud-news-app

I’d be interested in what you think about it. It’s completely open source, so any feedback or contribution is much appreciated! :)


I use Newsblur and see everything in the feeds I follow. The "training" and etc. are optional.


NB filters "inappropriate" images from my feeds, which is annoying.


Happy Newsblur subscriber here. I've never missed anything in a feed (other than the HTTP/HTTPS errors that my browser gives me)


Are you referring to it only keeping the last month or so of feeds?


Can you expand on this?


I'm working on this project, so slightly biased.

First, we show you both the chronological display in a list mode. That allows you to browse through things quickly and filter out what you're interested in manually. The personalization is used in combination with the image style layout. On that screen we try to surface the most interesting information.

Second, yesterday I was browsing quora questions about features of RSS readers. The most requested feature was personalization to help deal with information overload. I think some percentage of users doesn't need personalization, but the vast majority do.

Personally I was a big fan of Zite, I used that app everyday. Pity they were acquired and shutdown.


Zone is still around, buried in the options of Flipboard. The recommendations are weaker, but still valuable and worth the training investment.


I tried for a few weeks, but couldn't get Flipboard to show me content i'm actually interested in.


To get Cover Stories to be useful, it was necessary to downvote many of the initial selections, unfollow some tags, follow other tags, then upvote the few stories that were interesting. It gradually got better. But yes, there were many stories of poor relevance or quality. One problem is that the UI takes more clicks than Zite to rate stories and select tags.


You might like this:

https://www.goread.io https://github.com/mjibson/goread

Easy to self-host, solid replacement for Google Reader.


I'd love something where I can just cut all these guys out the loop.

I'd love an rss server I can run on my PC, it downloads all content for each new entry to my tablet at 6am each day. I can then read all content offline.


If that's what you want, I recommend tt-rss: https://tt-rss.org/gitlab/fox/tt-rss/wikis/home

It may be a bit overkill to run on your PC, but I don't know of a better program for that. I also sync it to my tablet every day for offline reading.


I came here checking to see if this could be a replacement for ttrss... But no. it appears that ttrss wins in the "just give me the rss feeds I want" category for me.


I’d recommend you too to take a look at Nextcloud News then: https://github.com/nextcloud/news/#nextcloud-news-app :)


Good to know! I use nextcloud, but it is performing TERRIBLY on my raspberry pi 3. I need to see about recreating the setup they use on the nextcloud box.

Is this better than ttrss? or just different? Worth the hassle of migrating things?


Interesting. How do you sync it to your tablet?

Do you get the full article each time for offline reading? If so what part is responsible for this? tt-rss or your client?


tt-rss has a cron that pulls the new feeds down, and then ties into a web-server that you can view the feeds on.

I have it sync to my android client via an app. I just put the web-server settings in, and it syncs.


Which android app? How do you deal with sites that serve up partial content in the feed, making you visit the site?


I use this one: https://play.google.com/store/apps/details?id=org.ttrssreade...

> How do you deal with sites that serve up partial content in the feed, making you visit the site?

I just don't look at those feeds until I'm online again. It's generally not an issue for me personally.


hmm. I think I need to abuse one of the real clients like feedly by giving it a url to my washed rss feed as this does offline content but doesn't deal with sites that serve partial content through rss.


I tell you what you've nailed it here. What I actually want is a podcast app but for articles.


I'm a fan of Tiny Tiny RSS (tt-rss). I self-host it at home. I use a mix of the web UI and the Android app to view it. I tried a few cloud-hosted readers but didn't like their UI as much as tt-rss.


Same. It's great.

Can even run it in a Docker container. https://github.com/linuxserver/docker-tt-rss


NetNewsWire has been doing that for me for years. I keep 30-40 feeds in a handful of folders. When one of them gets too spammy or boring, I delete it; when I find something potentially useful, I add it. I have never felt the need for some algorithm suggesting even more feeds to read.

I also write a weblog useful to a non-techie audience, and am disappointed and slightly frustrated by how few of those people use RSS. If you want to solve a problem related to RSS, solve that.


I've been a Feedly user since Google Reader shut down and have been pretty happy with it. The tings I subscribe to there I read from top to bottom every day.

When I'm looking for something outside of that, I've found the Apple News app to be shockingly good. For the things I read, it does a much better job of showing me interesting things than Google's Newsstand app. This surprises me because Google knows a lot more about me than Apple.


Indeed. I've collected a bunch of feeds I like to keep up to date on. What I really need in an RSS reader is the ability to import them (from Feedly if that helps) and then I want the feed reader to poll those feeds and let me read them. That's all I need. If I can chuck articles over to Instapaper et al, bonus.

Actually what I want is a podcast app but for articles. Let my device do the work.


A lot of the "recommendation" apps are underestimating human ability to comb through nonsense. On a social network like FB where primary use case is people and their personal stuff, recommendations of news items might make a lot of sense. For RSS readers where I have explicitly asked for what I want it makes no sense.


> Do people really want recommendations?

I do. Or rather, I want a service that will filter out the crap I don't want to see and bubble the relevant content to the top. In my job I have to keep an eye on loads of feeds, and it's almost always worthless dross, but I can't tell that it's worthless until I've seen it.


> but I can't tell that it's worthless until I've seen it.

So how do recommendations help?


I meant that of all the feeds, a small proportion of the content is valuable. I have to look at the non-valuable content to determine that it isn't valuable. I would like it if an automated system could score the content based on my criteria and behavior, reducing the amount of non-valuable content I have to filter manually.

As I understand it, that's the functionality Winds promises, but it only does it with a predetermined set of feeds, which isn't useful to me.


Newsblur has that feature. I don't use it, so I don't know how well it works.


Yeah, it works reasonably well, but you have to manually set filters and keywords, which is a drag.


Try http://news-ai.com/

May or may not work for you due to some bias in selection of feeds.


I'm with you. Feedly still does a decent job of letting me track the feeds I know want to track. Though I definitely don't check it as much as I used to.



I have a script that sends RSS feeds to my email and then filters that organize them as necessary.


Sorry I'm not interested in a Feed Reader that tries to give me suggestions. I like to follow the stuff I like to follow. I do the footwork of finding quality sources by myself.

Maybe I'm not most RSS users in that regard, I'm not sure, haven't done the market research.

It feels to me that suggested sources leads to promoted sources leads to paid content.


I'd actually bet this isn't going to be monetized that way since Stream is using the reader to showcase the personalization service they sell.


Not very fond of their "Try it out" feature, which after making me waste time on their "select 3" first step, then proceeds to try and cajole me into giving away my email address.

Wish they posted some screenshots or static pages or something of the sort, so I could take a look and see if I like it more than stringer, which is the current solution I'm using.


A little worse than that: Even once you do all that, the feed is empty. It seems to only get stories posted after your account creation, rather than pulling the existing feeds like most RSS readers.

So after doing all this setup, I had to go back to the site later to actually try out how it arranged content.


Yeah, a fledged out demo account is a "must have" when showcasing a product.


Some 500s due to HNews traffic, try the github repo: https://github.com/getstream/winds


Yep, that's when I left the funnel :) too much friction


In the category of self-hosted stream readers there's also Tiny Tiny RSS [0], which I use and enjoy. It has a nice Android app and pluggable scrapers for extracting webcomic images and full content from summary feeds. Although it's been described as having "the most hostile primary maintainer of a piece of software I've ever encountered", which I experienced a bit of personally last time I tried to contribute upstream.

[0] https://tt-rss.org


Did you try out Nextcloud News? https://github.com/nextcloud/news/#nextcloud-news-app

We have a more modern design than Tiny Tiny RSS, are way friendlier, and if you would contribute that’d be majorly awesome! :)


For work (writer) I go through dozens of feeds every day.

About 90% of what I see is irrelevant, but I need to see it to filter out the relevant data. I'd pay a decent amount for a service that can tailor itself and deliver relevant results without much manual intervention. InoReader and some other RSS readers have decent filtering, but it's all manual.

So this excited me, but without a way to import my existing feeds it's largely useless to me.


This is by far the most commonly requested feature. Top of our list for 0.2!


It's going to be extremely hard to convert anyone who uses an RSS reader without import functionality.

There is 0% chance that I'm going to switch no matter how good your recommendation algorithm is if I can't also keep up with my current feeds.


It wouldn't be so bad if they had like a suggestion box or something at the side with feeds you might like to add.



We've been taking suggestions on the GitHub Issues page. One of the biggest things folks are asking for is an OPML import.


I'm confused. It looks like this won't work unless you use the stream API, which is (apparently) not open source / self hosted.

The only reasons I can think of to host my own RSS reader are to prevent third party profiling, editorial transparency, and future proofing against the provider shutting down or evolving the service in a way that I don't like. This provides none of those advantages.

Am I missing something?



Looks really good! The design is fantastic, well done.

My only criticism would be that the Getting Started page is definitely skewed towards a specific segment. Like, how is Sports not an option? Music? I understand that there are only 9 slots in your designs, but please choose some that are more representative of general interests, not just Palo Alto mid-20s coffee shop startup founders.


We launched this on Product Hunt yesterday so it's currently tailored to that audience. More topics are coming! Going to need some community help with curating the feeds for those topics though.


Which features are you looking for in an RSS reader? Some ideas we're discussing for Winds 0.2.0:

  - Secondary links (ie comments link for HNews and Lobsters)
  - Batch OPML import
  - Deploy to Heroku button
  - Follow suggestions (we're working on this)
  - Switching between feeds should be easier
  - Lightweight task queuing system for emails and discover endpoint
  - Keyboard shortcuts (vim style)
  - GraphQL style APIs so you have more flexibility for building your own mobile apps
  - Android & iOS apps
  - Support more sites (RSS data quality is pretty poor and often needs custom logic per site/feed)
  - Search article's you've read using Algolia
  - Folders/Groups
  - Sharing support (e.g. Buffer, Facebook, Twitter, etc.)


hackuser posted a requirement which has my full support: Since you're doing machine learning, try and apply that to avoid showing me articles with overlapping content. Often quite a few sites I follow will cover the same thing, and obviously, I want to read it only once.


To clarify my version of the idea, I would like to see all the articles on the same subject but grouped together, perhaps in a collapsible outline, so I can choose which one or more articles I want to read on the subject.


Thank you, good suggestion!


What is batch OPML import? I'm confused by the use of the word batch in this context. How is this different from plain old OPML import? It is not like someone is going to import their feeds one at a time.

Anyways, any RSS reader lacking OPML import and export is a non-starter in my opinion, at least for me and anyone else who has a long list of subscriptions to setup.


I think, "batch OPML import" is just supposed to mean batch import of feeds via an OPML file...


Grab each URL and all referenced URLs with wget (as WARC) and optionally youtube-dl so I can have an archive of them all.


I was going to say mobile apps. Winds looks great--I just usually use RSS readers on my Android device.

Also, distributed sync would be awesome (i.e., without a centralized server, using something like syncthing or whatever), but that's just a dream.


Just in case you didn't see it:

https://news.ycombinator.com/item?id=12933006


Digg RSS Reader has been unbelievably good. The best thing about it (so far - knock on wood) that they don't tinker or change anything, because everything is working like it should. I don't really see a reason to try anyone else.


How are people still launching apps with sign-in over http - jesus christ.


Does anyone know an RSS feed reader with these features?

1) Deduplication

2) Grouping of articles on the same topic. For example, grouping all articles on Angela Merkel's latest Brexit comments, so I can easily choose one and ignore the rest.

Grouping would make me an order of magnitude more efficient when reading RSS. I'd happily pay $100 for that feature.

Also:

3) Efficiently and automatically manage broken feeds: Automatically retry (over X hours or days), automatically look for a replacement feed, and then let me know which feeds need my attention.


I run an aggregator using wordpress, feedwordpress plugin, and related post plugin - it does a good job of "grouping" articles of a similar topic. So I get a good idea of what other news sources has said about that topic. - it took 10 mins to setup.


I designed one a few years ago, but it was to help the newsroom of a magazine who wanted to publish blog articles every week. You can check it online, it was called "vendredi". It worked well.


Could you provide a link? I looked around and couldn't find it. Thanks. I'd love to try it.


https://fr.wikipedia.org/wiki/Vendredi_(hebdomadaire,_2008)

I am afraid there is not much left online now.


By deduplication, do you mean that you want articles on the same subject to be filtered out or do you just want the same article not to be displayed multiple times?


> By deduplication, do you mean that you want articles on the same subject to be filtered out or do you just want the same article not to be displayed multiple times?

I just want the same article not displayed multiple times. For example, if you subscribe to multiple feeds from a newspaper, some articles which fit more than one category will appear multiple times.

Articles on the same subject would be handled by 'grouping', according to what I wrote above.



Nick here from GetStream.io (Winds). We didn't expect Winds to see so much traffic today, so we're experiencing temporary downtime. We'll have everything back up and running as soon as possible. Thank you for all of your support!


Back up and running smoothly. Sorry for the delay!



Our team at Reamaze is all trying Winds right now since seeing on Product Hunt yesterday. Love the machine learning in-progress. I find the content to be curated quite well for my preference. The main issue I'm having is actually adding RSS feeds from sources I read on a daily basis. Not sure what the issue is. Love the two display options since not everything is relevant all the time and it's just easier to go chronological.


Just open a ticket on Github with the RSS feeds in question. We'll gladly add support.


since this is built on top of Getstream.io https://getstream.io/ I guess it follows their pricing structure - https://getstream.io/pricing/ free for 3 million updates...

what defines a feed update:

   The number of feed updates depends on how many 
   activities you publish via the API and the number of 
   users who follow those feeds. If you add an activity to 
   a feed that is followed by 50 other feeds, this will 
   trigger 51 feed updates. 
which sounds to me like if you use them to do stuff you have to control not just your feeds through their api but also your users - otherwise how will they know really how many users following the feeds?


The GUI is nice, very well done. It suggested a good website I didn't know, I will give it a try.

The thing I didn't liked is that, if it's supposed to be "learning" how comes I can't load new stories after I'm done with the initial set?


This website has absolutely no information about the product it's trying to display.


This is a bit too full on for me. I'm currently using feedly at the moment but I'm looking for something I can read without a browser. Does anyone have a recommendation for a battle tested rss reader with a good cli interface?


Does anyone have actual screenshots? I don't want to put my email address.


Looks like there are a few screenshots on the project site:

https://github.com/GetStream/Winds


It would be nice to be able to try the whole app without signing up with my email. That's what I thought "Try Winds RSS > Clicking 3 feeds > Continue Now" would lead to.


I use Feedly pretty much everyday. A recommendation engine for news doesn't really cut it for me. But of course, it would be a good exercise in ML.


It says Node 7, is there any feature that you use on Node 7 that is not available in previous versions? Just curious.


does it support firefox?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: