Grove is fantastic. I've been using IRC to communicate with my team forever and we were just using freenode. I finally got around to setting up an IRC server on one of our machines but when it came time to configuring and securing it properly it was going to take days to read through all that documentation, so I just locked it down at the firewall level and we all connect using a forwarded SSH port, or from the local host (we all use irssi as our chat client anyway).
A (rather large) ad agency that a friend of mine works for has a small group of people that wanted a collaboration solution and they all liked IRC so I sold them a monthly VPS with an IRC server configured but guess what - they have to forward a port to connect. That was cool, they said, they're on OSX and I provided them with scripts that just setup the port and opened Adium for them in a single click ... then along came a contractor that wanted to use windows :S
So I said look: you can pay me to figure out all this shit for you, it'll take me days and cost you thousands of dollars, but for that same money you could buy a few years of a grove.io subscription and I can still connect all the same bots that I've got running for you.
Seriously the price point they've got there is great. Properly configuring and securing a high availability IRC server isn't as simple as "set and forget" you have to do quite a bit of reading and configuring in order to get it running properly. Plus Grove has a bunch of web configuration options that are really handy for less technical folks.
IRC is the chat technology of the past, present, and future, keep it up guys.
I agree with the praise for IRC, but I don't see what you need grove for. There's mibbit, irccloud and a few other web IRC clients, all of which seem quite a bit more advanced than grove. And for the server-part; why not hop onto freenode, make a private channel - and off you go?
If you insist on private server then that shouldn't take someone linux-savvy more than 15 minutes either.
You literally type 'apt-get install inspircd', tweak two and a half values in the config file (to enable SSL and set a server-password) - and that's it. An irc-daemon is not exactly maintenance-heavy, mine has been running for over a year without a blip.
Firstly, you're not actually allowed to use freenode like that, technically it's supposed to be for open source projects right?
Secondly, I wouldn't run a server application like that on a publicly accessible server without learning pretty thoroughly about how to configure it securely, ymmv obviously but I just wouldn't do it.
Thirdly, there are a bunch of web configuration options that allow non-technical folks to login and do a bunch of stuff. If my client were paying me to do that for them they'd blow their Grove.io budget by an order of magnitude or two each month.
I signed up and tried grove a few weeks ago, I didn't really see the point. I then stumbled across some video tutorials for using hook.io to connect to an IRC server (#4,5 in the list of video tutorials) (http://blog.nodejitsu.com/hookio-video-tutorials) and immediately thought of grove.
It seems, to me, that they decided to use the tutorial, hook up web sockets (#9 in the tutorial list), add a nice interface, log all the events hook.io emits, and decided to charge way too much for it.
I could be missing something, but that was my initial impression.
Not a good example: Github thrives in part because it's a community, not something you can achieve if everyone is hosting their own git servers. They don't simply make an hard process easier.
The same logic could apply to literally any service. Restaurants? Why not grow/hunt and cook yourself? Translator? Why not just learn another language? Taxi? Why not buy and drive your own car?
For most of these things, the answer should be so obvious that it needn't be said, but apparently that's not the case. Division of labor and specialization generally has at least one of the following advantages over doing it all yourself: cheaper, faster, more reliable, more competent (i.e. better quality), and less stressful.
To be fair, unless you're a huge organisation with thousands of users, an IRC server is just a single process on a low-end machine or small VM. You can't really compare that with the whole infrastructure Heroku provides.
As a hook.io dev, this was my initial impression as well. However, it is much more difficult to develop this kind of product than you have described here. For example, the hook.io webserver was unusable until just about a month ago.
It's not like every YC startup was a paying customer of Convore. I'm pretty sure he meant that they used Convore instead of/as well as a mailing list to communicate with the rest of the batch. He was making a claim that Convore leads to better communication than a mailing list, and that in the case of that YC batch, it was enough better that he thinks it made the companies noticeably better overall.
But 200 people using it for free (and probably naturally fading out of using it once the summer is over anyway) is not going to keep them afloat.
"What actually made the startups switch away from convore?"
I don't think anyone switched away from convore as much as they just stopped using it over time. The problem with convore was that it was too focused on information, and people don't care about information, they care about people. So what started as a better version of chat/forums eventually devolved into a less good version of quora.
The best thing that convore had going for it was that all of the content was extremely lightweight, but in the end I think this is what killed it also because the emotional connection just wasn't there.
Their new startup is Grove, which is hosted IRC for teams. I'm guessing this is what most of the YC batch switched to to replace Convore, since it hits more directly on their need.
While I'm not a fan of Leah Culver at all (I view her as Eric Raymond^2, better at selling the image of geekiness than at actual coding), most startups fail even with good founders.
Besides, don't we pride ourselves on accepting failure and treating it as a learning experience?
I've met her a couple times, and have a different opinion. (I admit, I was biased a bit before meeting her, based on all the press and the murder of a friend's RAID while drunk...)
While she might not be the best computer scientist in the world of startups, she's great at the other roles in a startup, and is certainly a competent developer now. Not everyone in a startup needs to be a classically trained computer scientist with a Stanford PhD. Not every successful startup has to look like Google. You know who also wasn't a great computer scientist? Steve Jobs.
(I actually met esr when I was growing up in Pennsylvania, too -- he ran the local community Free-Net which I used for Internet access. He was a hardcore geek back then, and seems to have shifted more toward self-promotion and politics over time. He actually wrote some decent code for BSD/OS and ISP operations.)
Ellie, don't begrudge someone because of misinformation. Sixapart acquired Pownce, and Convore is a success since the fruits of the developers labor live on in Grove. I still remember Leafy Chat, which is probably the inspiration for Convore. Regardless, they all were great products.
Powce was successfully sold to Six Apart. She has a proven track record of building things that people like and want to use, but you can't win them all.
Sure. They went straight to work on something similar that Six Apart was cooking up. But, there have been many amazing products shutdown over the years due to talent acquisitions.
Pownce ultimately would have been forced to shutdown due to the popularity of Twitter and Facebook I think, but that doesn't mean that it itself wasn't a good product.
Late last year I was in an incubator where Convore started. There were 2 other companies doing the same exact thing .. All of the YC companies were using HipChat, Convore seemed like "yet another copycat site to sign up for an account with". Chat is a solved problem. Let's move on.
Nothing is a solved problem. HipChat's a great product, and in my opinion the current market leader, but there are always ways to innovate and disrupt. (That being said, using IRC probably isn't it.)
No, I do mean it literally. The wheel is a pretty basic invention, and yet tire manufacturing is a $140 billion industry that spends a great deal on research and development.
When you think a problem is solved, it's actually in a perfect position to be disrupted.
Yes. My point was that there's a huge amount of effort currently being expended to do just that. No one is trying to change the fact that a wheel is round and rolls, but from there, there's quite a bit of space to innovate.
Agreed. But their "pivot" into Grove is actually really valuable. Chat is pretty essential for teams (ours at least), but it's not that easy to build it yourself without running into major scope creep and other headaches. I'd rather pay a little for reliability rather than distract the team from the main goal to build a chat system.
I would have preferred to see an extended read-only period, rather than a month in which to panic-download your data (and the export functionality seems to be broken right now too https://convore.com/export ).
I'd really like it if there was more time to capture the site and submit copy to the internet archive. I'll see what I can do with archive team.
So long, and thanks for all the data! Perhaps convore is an apt name for something that consumes your history. :-)
Export worked for me, but it's moronic. It seems like every user gets all his utterances.. which are utterly useless without context.
They're basically treating the data as if it's email or facebook posts and comments (those last also mostly useless without context). But it's not, it's a far finer granularity of interaction.
Convore, please let users download logs for the channels they're members of.
I'm working on a site with functionality similar to convore. Discussion is centered around links in a manner similar to reddit or HN. Just finished an alpha test, you can register for the beta at http://b00st.com.
we've been using campfire forever (and I, for one, am super happy to see competition. will definitely be trying this out). we want the functionality of IRC, but we dont want to deal with setup, maintenance, security, etc.
its the standard SaaS argument: the base case might seem easy, but the details are going to drive you to the bottle.
Especially since Skype gives you all of their functionality for free. I'm really wondering what advantage this has over Skype group chats (which is free).
Without even looking at their page: logging, persistence, availability, API hooks.
Skype is awesome and is the cornerstone of my team for private communications. But if history is only transferred when parties are connected, and only between people who are explicitly added to each chat, it leaves a giant gaping hole in group communications.
Availability means it's online all the time. A server isn't, people's Skype isn't. Maybe it works for you if you're all in the same office, but I have people in 8 different time zones.
Internal to a team, not for large-scale public servers. $10-20/mo to have everyone on your team have IRC without having to manage it, and with features rolled out over time (disconnected mode is nice, and search, and maybe bots) seems totally reasonable.
It's just that setting up IRC is so trivial: apt-get install inspircd. I have been running an IRC server (a reasonably public one, even) since 1994, and even at $250/hour (the highest hourly wage I've seen people reasonably claim for technology work) I don't spend $10/month administering it (and remember: that's just for 5 users...).
Of course, you need a machine (virtual or otherwise) in the office for that, which will always be online. And you can't just add outside contractors to your chat server if they're not in-office, or connect up yourself if you're on the road.
You also don't get searchable history (archives) by default, which is a big win in a lot of situations.
I'm an experienced sysadmin and developer, so managing an IRC daemon doesn't scare me, yet I'd pay $10 per month for all of that in a heartbeat.
Ok, if you are a company that doesn't have any machines in the office that are always online, I'll buy it. I was about to wonder "where did you find such a workplace...", but I failed to take into consideration non-technical companies.
(That said, again: it isn't $10/month, it is $2/user, and the /user price increases with more users. Even for my relatively tiny company, it would be more than $10/month.)
Im glad the team is moving on, bigger and better! But why not keep the site up. How expensive could it be to show your commitment to the few users you do have. If anything it builds a good reputation on keeping your projects going even if it wasn't the massive success the founders wanted. Im all for progress but every project doesn't need to die does it?
I still get a weekly email from Convore telling me about new posts in the groups I belong to, and when I try to click through to them, I get sent to arbitrary pages. How can one unsubscribe from the mails?
It's not really about Grove vs HipChat but more IRC vs HipChat. IRC is a really old and open protocol, meaning that hackers built lots of tool around it and already love it. But it's a pain to configure and maintain it if you've got more important things to do. (Like building your app). A couple years ago, I believe grove wouldn't have stand a chance against the hacker mentality of "I'll do it myself", but now that more and more companies need to go as fast as possible, it makes sense to pay a little bit to have a service that's stable and just work out of the box.
And, for me personally, I hate HipChat clients (IIRC they run on a crap adobe version?) and I like using the existing tools I have for IRC. For instance, I've already setup everything to work on my vps with irssi, with the plugins I want, the theme I want, etc. I don't want a "new and prettier" solution to a solved problem.
I'm not a fan of HipChat's official native clients (well, as 'native' as you consider Adobe Air), but the web client is very nice and (more importantly) it's trivial to get your HipChat account set up in Adium/Pidgin/<insert your favorite IM client with Jabber support>.
I'd say it's about "the chat tool you use" vs grove. In my case, it's also HipChat/Campfire.
Unless you're focusing SPECIFICALLY on "hackers that love IRC" - in which case your market is quite small (compared to the whole corporate chat market), imho...
Well, "hackers that love IRC" is still a huge market if you think about all the software companies. And, even if it's a smaller one, sometime it's better to create a product that'll be used by less people rather than trying to build something for everyone and ended up with nobody using it. But I feel your concern, Grove might become a successful small business but they'll have to attack other market or build different products to continue to grow.
I'm not trying to dance on their grave or anything, but from reading this article it appears a company I worked with last summer had read it as well, and that level of -- let's call it "sophisticated" -- messaging architecture was beyond unwieldy for an early-stage company. I can understand wanting to avoid Michael Arrington calling you "amateur hour" when things fall over under popularity, but there's a such thing as overengineering.
I realize my opinion might be based on a red herring or strawman, but when chat is the "Hello World" of server-side JS I'm just not sure there was enough outside of the architecture to hang a company on.
I never quite got Convore. Grove is sort of interesting, but I doubt it'll displace us from HipChat.
The major advantage of using a non-proprietary protocol like IRC is that people can choose their own clients... but the state of most IRC clients nowadays is abysmal. Colloquy on OS X, for example, hasn't been updated in years.
If I was going to use IRC to communicate with my team, I'd just create a private channel on Freenode. We'd miss out on archiving, but I don't think that that feature alone is worth $2/user/month.
Freenode is for free software. It's not "free as in someone else runs our private chat for us." Hosting business chat via terms of service abuse is inadvisable to say the least.
At no cost you could connect to the channel with IRCII (or pick any other logging bot/client), activate logging and write a shell script to pull the logs into a website at midnight every day.
Colloquy can be pretty bad, but I've been satisfied using either Adium's IRC support (though I wish I could easily block the server message / nickserv spam) or LimeChat, which was the simplest/fastest standalone OSX IRC client at the time I was using a standalone client.
With Adium I've been pretty happy since I discovered the ability to turn off show/join messages on the history menu.
The other thing was the Renkoo message style (from preferences, messages) was much better than what I had before. Date and name to the sides, content in the middle.
I have a Mac mini logged in permanently so I can review logs and there I took the advice which was to right click and block chanserv and nickserv. Not quite brave enough to do that on my laptop.
Convore was a pretty big hit at PyCon 2011. It was useful because it allowed people to create different rooms for each event going on at the conference. It worked really, really well.
Too bad it wasn't a product that could be sold. The technology and site were cool and worked well.
I really enjoyed Convore on PyCon last year. Much more than Twitter for example. I remember there were rooms on just about anything, and among other things I met some guys to grab beers with on Convore. Too bad it didn't work out.
The code would need a little bit of cleaning up before it could be released, but this sounds like a good idea to me. I left the company quite a while ago, so my say no longer counts, but I would be supportive if Leah and Jori decided to go down this route.
You're actually charging more per user for more users ($2 per user for 5 users; $2.50 per user for 50 users), rather than giving larger accounts a discount. I imagine the reason for it is the load it places on your infrastructure, but I've rarely seen that kind of price stacking work in the wild.
Past the first plan, the pricing is all $2.50 per user. I like to think of the first plan as the "startup" plan so we made it cheaper. We do offer discounts per user past 50 users.
I still dont get who would use grove? IRC users aren't exactly looking for a client replacement. Existing IRC clients just work, like Craiglist. Colloquy on the Mac, etc.
A (rather large) ad agency that a friend of mine works for has a small group of people that wanted a collaboration solution and they all liked IRC so I sold them a monthly VPS with an IRC server configured but guess what - they have to forward a port to connect. That was cool, they said, they're on OSX and I provided them with scripts that just setup the port and opened Adium for them in a single click ... then along came a contractor that wanted to use windows :S
So I said look: you can pay me to figure out all this shit for you, it'll take me days and cost you thousands of dollars, but for that same money you could buy a few years of a grove.io subscription and I can still connect all the same bots that I've got running for you.
Seriously the price point they've got there is great. Properly configuring and securing a high availability IRC server isn't as simple as "set and forget" you have to do quite a bit of reading and configuring in order to get it running properly. Plus Grove has a bunch of web configuration options that are really handy for less technical folks.
IRC is the chat technology of the past, present, and future, keep it up guys.