Hacker News new | comments | show | ask | jobs | submit login
F5Bot: Email Keyword Alerts from Reddit, Hacker News, or Lobsters (f5bot.com)
78 points by dustingetz 61 days ago | hide | past | web | favorite | 45 comments



I wish services like this would move beyond offering everything for free. It's a valuable enough service to charge a nominal fee, be it lifetime or whatever to ensure it sticks around, and give the developer incentive to improve upon it.

I'm guessing the next big web movement will be more paid services that actually have value. Weeding out all the trash that disappears with your data in a year.

Update: Looking at his Patreon. He's making a total of $41. Not even enough to cover his hosting fees apparently. That's why you charge for something worth charging for. https://www.patreon.com/f5bot


That's not how new services or startups work nowadays. You need critical mass first, and if you put up barriers - like paying money, like ads - people won't use your service. And there'll be a free alternative that takes over.

No, the formula is to invest your own money, hopefully see a lot of growth, get a few rounds of dank investor money to pay for it all and stay afloat, and maybe, MAYBE, come up with a monetization scheme (like idk, reddit gold or premium subscriptions), and then either get bought by one of the big companies or go to the stock market.


That's how yesterday worked. Impeding recession is going to dry up that funny money. You're going to have to actually build products people want to pay for to survive. Two individuals being equal, there is no way the one working for completely free is going to outpace the one getting paid to do it. That free service is attractive at first, until you're missing features, support, and longevity that the paid service has.

Example. pinboard.in


Pinboard employs a whole one person. It's a good model, but its aims are pretty different from those of your average startup.


There's two, and the average startup could learn a lot from how Pinboard is run.


I don't think this matches observational evidence. On the contrary, people building free services do that to maximize growth, which maximizes VC money, which gives them more resources to add features & support... until the service implodes or gets acquihired, that is.


You're talking about such a small subset of high profile sites, that are living in the SV bubble world. There are thousands, and thousands of other services out there that people pay for, that Hacker News has never heard of, and have no intention of burning VC money or playing the acquihiring friends game.


The alternative, which I find really attractive is apparently the model that WhatsApp used. It was free for at least a year, after which time, one was asked to pay a small fee. If you use it for a year, you are a finding value, and will most likely convert. Either charge a lifetime or yearly fee.


I am working on a similar product, https://www.pagespy.io, that lets you monitor any webpage, albeit not as simple and targeted as f5bot (I really like the name), currently in very closed beta, there's definitely interest for such applications, currently it pays for itself and some ramen on top.


Nice! Congrats on shipping. Don't be ashamed to charge for a good product. Your potential customers will decide if something is worth paying for. The era of growth over revenue is coming to an end, along with an impeding recession!


Thanks! It only took me 11 years to get the courage to at least _attempt_ to ship something. And although I am no position to look for investment, as in I don't have access to a VC/investment environment, I can dedicate at least 2-3 years to the product, bootstrapping it.


The next best option is to open source it or package it in a way you can run it at home forever. I actually have a very similar NodeJS scrip that does the exact same thing on my GitHub for 100% free. It uses Amazon AWS SNS to send the notification emails or text messages. Only supports Reddit ATM.

https://github.com/VictorioBerra/reddit-aws-notifier


Well they are charging and only taking in $41/mo, so perhaps this isn't something that many folks are willing to pay for.


They aren't charging, it's free. The Patreon is donations. Put up a paywall of $1 a month, and I GUARANTEE you will make more than $41 a month.


Patreon is nominally about donations, but depending on your level of support, you get different "rewards" for your money. I believe that they were originally for just providing swag - Send a creator a few bucks, and they'll send you a few stickers, send them a bunch and they'll send you an autographed piece. But many Patreon pages offer real features at certain tiers: real products as rewards, exclusive access to materials, or, in this case, exclusive functions from the service.

> Power User

> $7 or more per month ∙ 0 of 25 patrons

> As a Power User, you'll get access to exclusive features such as the ability to monitor for common keywords and a higher daily hit limit. I'm actively working on adding new features too.

It's like a subscription, but with slightly less of a service guarantee. It's just a convenient way to get paid. The page does say "The service is completely free. There is no artificially low usage limit, and there is no up-sell." and he could do a better job of monetizing it by, say, limiting each account to 50 free notifications and requiring registration of a new account after that or offering an up-sell to a $1 paid tier.

If you want to let some people use it casually for free, it's really hard to charge others who would be willing to pay more.


Sure, they may make less money, but your major gripes with free services:

"It's a valuable enough service to charge a nominal fee, be it lifetime or whatever to ensure it sticks around, and give the developer incentive to improve upon it."

These will all be resolved once the Patreon hits the $100/mo mark, which I GUARANTEE will be reached pretty quickly if it is indeed a valuable enough service.


Considering that's not even identified on the front page of f5bot, doubtful. People are more apt to pay for something, when they have no choice.


> and I GUARANTEE

Only if people think it is worth it. $12/year for notifications is a bit much, IMHO.


Depends.

I actually started writing a similar project on the side for HN only, just because I hope getting notifications for a bunch of topics will reduce the amount of time spent refreshing the frontpage here. This translates to life satisfaction and sometimes real money (procrastination is a bitch when you're freelancing). I'm going to test this particular project now, instead of finishing my own, and if it works well enough (over the next couple days/weeks), I am going to pay it $10/month.


Keyword alerts are interesting, and it would be pretty easy and relatively cheap to do.

I've already built a little app with a Telegram interface that allows you to subscribe to users' comments and posts. Every x minutes you get a message from the bot with the post titles or comment bodies for the users you follow.

I'm curious if it might be useful to expand my tool to be able to track keywords, and if there'd be enough interest for me to make it public-facing on some server.

But for now it's proven surprisingly useful. I check HN less and just click on the links that arrive in my (muted) channel (I got here because I followed you, actually!)

The biggest annoyance right now is that I regularly get an update that will contain a number of comments by a number of users I follow, but all in the same thread. it's a bit annoying that I have to process a bunch extra requests to get to the root post (at least with the official API?).


> I got here because I followed you, actually!

Huh. I'm flattered :).

In a way, I think there's a triad of required features for productive use of discussion groups: following topics that interest you, following people whose comments you like, and getting notifications when someone talks to you. So following users and keywords are two big things missing for me (notifications on replies I get through HN Notify http://www.hnreplies.com/ by 'dangrossman). I find that I often miss discussions on exactly the topics that interest me, because for some reason they don't show up on the frontpage at the time I'm looking; I hope a keyword-based search will plug that hole, and decrease the urge to refresh the frontpage so often.

A keyword-based system doesn't have to use e-mail - in fact, I think it would be better as a dashboard (the side project was working on was actually going to be a simple webpage with a list of matches). The benefits of such solution would be that a) you can easily de-duplicate matches, solving your biggest annoyance and avoiding spamming someone's inbox, and b) reducing the distraction factor a bit more for people who didn't yet decide to disable notifications on their e-mails.

That Telegram tool of yours - any chance I could take it for a test drive?


> > I got here because I followed you, actually! > Huh. I'm flattered :).

Well, you're quite prolific and the signal to noise ratio is pretty decent :). Speaking of: it's much more interesting than I thought even just to see the frequency, time, and interactions of users on HN via Telegram updates. For example, I've found that some users seem to be active at the same times, and (as a result?) get into discussions with each other regularly. I would've never noticed just by my (compulsive) regular browsing.

> In a way, I think there's a triad of required features for productive use of discussion groups: following topics that interest you, following people whose comments you like, and getting notifications when someone talks to you. So following users and keywords are two big things missing for me (notifications on replies I get through HN Notify http://www.hnreplies.com/ by 'dangrossman). I find that I often miss discussions on exactly the topics that interest me, because for some reason they don't show up on the frontpage at the time I'm looking; I hope a keyword-based search will plug that hole, and decrease the urge to refresh the frontpage so often.

Yeah, good observation. I've concluded (and experienced) the same. I chose following users as a first itch to scratch because it seemed easiest and most useful at first. notifications when responded/talked to would be a close second (but I also use hnreplies.com for that; thanks Dan!). The keyword one I'm a bit more uncertain about. I use hn.algolia.com quite regularly for a few keyword searches, and I suspect that 1) there's a bit too much 'noise' for many keywords, and 2) it's quite a bit heavier to implement. I think ideally it'd be a bit 'smarter' somehow than just keywords (even if just AND/OR/WITHOUT, but ideally taking into account upvotes or replies). But how much of a problem this is depends on the channel used. For example, having a dedicated Telegram conversation for a keyword would probably be okay. aggregating this stuff in an email less so, at least not without more contextual information.

> A keyword-based system doesn't have to use e-mail - in fact, I think it would be better as a dashboard (the side project was working on was actually going to be a simple webpage with a list of matches). The benefits of such solution would be that a) you can easily de-duplicate matches, solving your biggest annoyance and avoiding spamming someone's inbox, and b) reducing the distraction factor a bit more for people who didn't yet decide to disable notifications on their e-mails.

Yeah, agreed. The way I see it, the tool I'd like to build or pay for would allow you to pick 1) a source website/page, 2) specify this a bit (elements on page, number of updates for source, whatever is appropriate), 3) specify update frequency, and 4) specify the channel for updates (email, telegram/IM, RSS reader-style interface).

Ideally said tool would be able to suggest elements to 'target', or offer to follow the RSS feed instead (when available).

> That Telegram tool of yours - any chance I could take it for a test drive?

It's an unholy mess of Elixir code right now, and I've never put any of my stuff on github yet, so that's a bit of a barrier, but I'd be happy to let you know if I've got it in some kind of shape that I'd be comfortable sharing :). I'm of half a mind to expand it anyways.

That said, it's pretty simple. Checking for new user posts is trivial, as is sending a Telegram message (just a GET request). most of the effort would be to make this work for multiple consumers.


> it's much more interesting than I thought even just to see the frequency, time, and interactions of users on HN via Telegram updates.

Uhm, so now you're aware just how hard I procrastinate :<.

> For example, I've found that some users seem to be active at the same times, and (as a result?) get into discussions with each other regularly.

That's very true. I also wonder if this isn't partly responsible for differing perspectives HNers sometimes have about HN. I sometimes see people claiming that HN is full of things I've never really seen here. Maybe it's just timezone difference? Maybe HN does look different at +12h offset?

> there's a bit too much 'noise' for many keywords

That's true, but I think just making the user responsible for picking right keywords would give 80% of the value.

Aggregated keyword matches is something I've never seen done in the context of HN/Reddit, but I feel it's worth testing. Maybe I'll finish up my dashboard, free time allowing.

> but I'd be happy to let you know if I've got it in some kind of shape that I'd be comfortable sharing :)

Please do. Don't worry about adding new features first; from what you describe it's quite useful already.


Couldn't build it yourself for less than $100. $12/yr is a good deal


Micropayments, crowdsourced


The author of F5Bot recently published a guest-post on the Intoli Blog called How F5Bot Slurps All of Reddit [1]. It goes into a lot of technical detail about how the service works and some of the challenges that arise when dealing with such a large number of posts and comments. If you're a fan of F5Bot, or just learning about it now, then you might be interested to check that out.

As others have mentioned, F5Bot is free and completely driven by donations and Lewis Van Winkle's generosity. Please consider giving a donation if you're a user of the service and find it helpful [2].

[1] - https://intoli.com/blog/f5bot/

[2] - https://www.patreon.com/f5bot


Thank you for the intoli link. It was a good read, plus I learned about Aho–Corasick algorithm


Ah, so that's how extremists seemingly always find Reddit threads related to certain hot topics and brigade them.

I used to think that it was a simple case of fringe folks being attracted to popular threads. But it happens frequently enough on smaller threads on otherwise moderate, modest communities that I thought something was up.


There is also this: http://wiki.project-pm.org/wiki/Persona_Management

And this: https://www.nytimes.com/2018/10/08/us/politics/rick-gates-ps...

I left reddit because with the API these disinformation campaigns are too easy to create. After watching from afar during the 2016 elections noticing weird new things happening, new patterns, I decided that I no longer knew what was real and left.


I've been using this for a few months and it's been really useful. I agree that offering a free version and a "premium" version with a few extra bells and whistles would be a really good idea. It's definitely the kind of thing that I'd pay to have extra features for.


Hi, F5Bot author here. I'm working on a premium version, and a couple features are already available through Patreon. It has just been a low priority. If you have ideas for any premium features you would like to see, please feel free to drop me an email sometime (just hit reply on any alert you get).


What’s the difference to a Google Alert with: "site:reddit.com KEYWORD"?


Do Google Alerts even work reliably these days? I set up a bunch some years ago, realized they don't seem to be working, and promptly forgot about this feature. Every month or free some alert I set up then triggers and send me an e-mail, but I kind of expected to get such e-mails every day.


They are hit and miss. Nearly 100% of the results are from news sites and it seems to lag. Like, I'll get forwards from associates about Topic on Monday then the Google alert on Wednesday


Google Alerts seems to miss most of Reddit these days. It's really broken.


Is it worth seeking out an invite to Lobsters am I better off lurking?


Lobsters is very low traffic compared to HN. The posts tend to have ~50% cross-over rate with HN as well.

anyhoo, I can hook you up if you want. DM/Mail/Smoke-Signals me.


This is the first time I've heard of it and this is awesome and amazing, and I'm so glad it is free. Thank you for making this available as a great tool!


I use F5 and it's been genuinely useful seeing where my own blog posts come up! Thanks to the creator for making it


Does the author have a timeline for when they'll open source it?

https://github.com/codeplea/f5bot


The about page says:

  If I reach $500/month in donations, I will open-source F5Bot.


Ah, that's fair. I didn't see that.


So what happens when people stop paying the Patreon and drops below $500, they close the source?


I built a similar service, but for podcasts: https://www.listennotes.com/alerts/


I always forget about lobste.rs I remember wanting a membership when it launched and getting mad that I couldn't find anyone to send me an invite. Kinda made me feel like it was the no homers club.




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

Search: