Hacker News new | past | comments | ask | show | jobs | submit login
Grove.io: Hosted, Searchable IRC Chat For Teams (readwriteweb.com)
227 points by jzb on Nov 8, 2011 | hide | past | favorite | 101 comments



I know it's brand new and all, but there seems to be a lack of "normal" irc functionality/usability. Nickserv has no 'help' functionality, nicks apparently have to be universally unique across all of Grove (better hope someone doesn't use my name in a different organization?), there's no chanserv. The help page is concise but should probably be printed in MOTD or some other way for non-Web clients (I was confused that I had to identify before joining a channel... it created a 'ghost' channel for me instead of booting me and telling me to identify, like most irc servers). And when you add a member to the org, they immediately are joined to the channel - even if they're not connected to a client? How do you know they're available to respond to a question?

Also, since it appears to be a new & custom irc server, i'd not really be willing to trust corporate communication on a platform known for attracting hackers.


You're correct - it's a totally new and custom IRC server (for good and bad). We built it as a minimum viable product are continually adding features - both standard IRC features and custom features for our customers.

Thanks for reporting these specific issues. We'll work on getting them fixed.


there seems to be a lack of "normal" irc functionality/usability. Nickserv has no 'help' functionality, there's no chanserv

This may shock and surprise you, but I've been using IRC on servers without NickServ and ChanServ for decades. No services != not IRC.

Also, since it appears to be a new & custom irc server, i'd not really be willing to trust corporate communication on a platform known for attracting hackers.

Obvious troll.


Granted, services are not required. But if you're going to charge for a hosted IRC server (and you require nickserv to join a channel) you might as well clean them up and make them slightly user-friendly/functional.

Rewriting an irc daemon from scratch is a bad idea. The multitude of existing irc daemons and their hybrids are proof that the technology's complex enough to take years to stabilize, much less secure. If it does actually succeed they'll run into the growing pains of scaling that (assuming they add standard irc cluster topology) will bring on netsplits and other potential issues. If I were a hacker i'd look at a from-scratch not-quite-complete alpha-release IRC daemon as a fun toy to mess with. And if I owned a company I wouldn't want to risk my private communications on such shaky ground.

I'm not saying it isn't a cool idea... it just needs a few years to cook before i'd put my money or business into it (or at least a reputable, secure codebase)


> Also, since it appears to be a new & custom irc server, i'd not really be willing to trust corporate communication on a platform known for attracting hackers. > Obvious troll.

I wonder why did the popular media portray IRC as a hacker hangout, when yahoo messenger rooms and other group chats are implementation of the same concept.


I don't think it is supposed to "be IRC". It's a chatroom backed by IRC, and it seems to work perfectly.


If it's not supposed to "be IRC" perhaps they shouldn't include the word IRC 12 times in the article, their website shouldn't have an "IRC Help" page and instructions to "set up your IRC client" and your "IRC settings", the front page shouldn't say "Hosted IRC" in huge bold letters and a "Sign up for your IRC server" link at the bottom. Might make people think it was IRC. But that's just my opinion.


"be IRC" == be a standard/traditional IRC server. Why would anyone launch that? I think you are missing the point.


Nice work, thanks for making pretty things, keep it up.

Disclaimer: I've been brainwashed by working in "enterprise" environments. I understand that is not necessarily your target audience.

Logging scares me when it is a hosted solution like this. I signed up for the trial (for personal use) and didn't see that configuration option. Maildrops are lots of fun, but dumps of your dev IRC chats? Most regulated environments have retention periods, and after that most communication is considered to be a liability and storage concern (IANL, TINLA). I would say that dev chats are useful for about a few months, max.

A non-hosted solution would be nice, but certainly takes a huge amount of effort to layer on something like that to what you already have.

Is there any open community (i.e. a grove.io instance) to join and chat about your product, get a feel for it? Sounds like a bit of maintenance headache, but probably worth it. I can't really demo it well by myself.

Also, I can't imagine talking to my coworkers like I am accustomed to chatting on IRC <ahem>. That's just a random thought.

Edit: spelling errours


Just saw a demo of this; it blew me away with a clean UI, and some awesome features -- the search/ACL implementations are REALLY slick. It also gets rid of another problem I have, which is that I hate installing/configuring IRC servers.

Every company I've ever worked for has needed this.


Not quite as shiny, but I've hosted IRC logs for a few years now. Example: http://irclog.perlgeek.de/perl6/today

If anybody wants a channel on freenode logged publicly, feel free to contact me.

(The search facility currently sucks, I'm working on a very different search backend these days).


What's the advantage of an IRC-backed chat vs. a Jabber-backed chat?

Sure, IRC's protocol is simpler, but it's also not designed to support authentication or security natively, or to be federated in a "cross-network" way like XMPP is. It seems that XMPP is more modern, more flexible, and supports real user authentication natively - a win in every respect.

Plus most of the clients people these days use for IRC (Adium, irssi, Pidgin, XChat) support Jabber anyway.


It's not? How about:

* passwords required to connect to the IRC server,

* passwords required to use given nickname (enforced via NickServ),

* SSL.


NickServ and "IRC services" in general are shoehorned extensions to IRC. Most clients support them in some automated fashion now, but there's no guarantee, and usage differs depending on "IRC services" package.

At any rate, my original comment came off a bit the wrong way. I'm not trying to say that IRC isn't a workable solution for this kind of group chat system, as it clearly is. I just don't see why using IRC over Jabber (which seems more technically suited to this domain) is a selling point, especially in a marketplace crowded with group chat solutions.


I think it's actually an issue of perception. When I hear Jabber or XMPP, I think of one-to-one chat and when I hear IRC I definitely think of group chat. In general, IRC clients seem more tailored towards groups.

That said, there's definitely room for both protocols and I could see supporting XMPP as well at some point.


XMPP MUC is not fun, nor fun to scale. IRC has netsplits, but it does work reliably for tons of people all day every day.


I don't really think that's true. Look at Jabber.org which is probably the biggest jabber network out there and it seems they're scaling pretty well.

Edit: Also XMPP supports UTF-8 for everyhting out of the box and as far as I know it's not really possible with IRC (or did that change over the last years?). E.g. nicks with non-ascii chars are no problem to have..

Why do you think it's not fun?


XMPP MUC also has the problem that the largest XMPP client, Google Talk in Gmail, afaik does not support it. It handles multi-user chats via some proprietary method that only works for Gtalk clients.


I'd be interested to know how this stacks up against HipChat. I'm not sure what I'm missing out on, but it would be good to hear if there is something.


The major advantage I see with this over HipChat is that you can use existing IRC client software to connect. Personally, I'd much prefer Irssi in a terminal to an Adobe AIR product.


You can connect to HipChat from any Jabber client, though unfortunately it makes chat rooms slightly less discoverable. (I’ve used Adium with reasonable success.) The big advantage is that (at least on a Mac) the Adobe AIR version is a tremendous CPU hog.


I can see the reasons for that, but the HipChat application lets you drag and drop files to share them with everyone in a room, auto-loads images... that alone is worth the client app to me.


You can do that with HTML5 (I suspect it's coming soon)


You can do that with the IRC client.


One functionality I miss from HipChat that I haven't found here, is the ability to copy files from your disk and paste them to the chat (or drag and drop), where they get automatically uploaded so all the chat members can view or download them. If Grove gets that, I'm sold.


Unless I'm missing some detail of what you're saying, HipChat already does that. I regularly drag screenshots into a conversation. Hipchat uploads them to S3.

There is a size restriction, though, which I've run up against more times than I can count, since I never seem to be able to remember that it's there.


What @napoleoncomplex meant was hipchat has the file upload feature, but grove doesn't. And he would move to grove if that was present.


that i can use any irc client i want and not install adobe air or use their web-based chat


You can use XMPP clients with HipChat.


Awesome, this service is needed. I had the job of finding/implementing an outside-accessible group chat service for a startup. I wanted Campfire, but one of the founders didn't like newfangled web chat and required it be traditional IRC.

Setting up open source IRC servers is a pain. They are almost completely void of documentation. And as the article points out, IRC doesn't usually have the traditional username/passwords or access controls you really need for business group chat.

I realize this is an MVP. Here are somethings I think should be included:

* Groups. It makes it way easier to manage access controls.

* A way to invite outsiders. The specific use case is inviting customers to join a chat, without the customer taking up a user/license seat.

* You've probably done these but I can't tell from the website; persistent chat rooms and varying admin privileges (ops in some rooms, not others, ability to create rooms, etc.).

EDIT: Why was my comment down voted?


This would be cool if it included what I think is the most useful aspect of IRC for teams: trainable IRC bots.

Back in 1999 - I used IRC to manage my IT org. We had two IT channels: one for IT staff, which we were all required to be logged into when on the clock - and another that was company wide, where users could jump on IRC to ask the IT/dev/support teams questions.

We used bots to train IT FAQ and system info...

So for example, we would type "DNS Server" and the bot would reply with the IP of the DNS server.

We could type in a hostname and the bot would reply with the system details we put in there.

So, i this had the capability of learning and storing data that was retrievable in this way, rather than simple search, I think it could be very useful.

User:"Whats the current version of ProductX"

Bot:"Product X current version is 1.1.2"


Hubot has this feature, and it also has an IRC adapter.


Hmm, I just started working on an open source web-based IRC client a couple of weeks ago. This takes a bit of the wind out of my sails.


Grove is primarily an IRC server for your business (the client is tangential), so please don't give up! Actually, feel free to send me an email - I'd love to help you get your web client working smoothly with the Grove server.


Something that works with existing irc servers would still be fantastic.


Yeah, the cient looks pretty fantastic, but the apparent limitation of working with Grove-hosted servers does limit its usefulness.

If you're interested, here's my client so far:

http://web-irc.nodester.com/

https://github.com/akavlie/web-irc


http://usealice.org works really well.


Don't let it get you too discouraged- Leah doesn't last very long with any one project so you'll be fine in a few months.


Even if this is true (I have no idea whether it is), making this comment here and at this time is in extremely poor form. You should be ashamed of your utter lack of simple etiquette.


He's at -9 karma, I imagine he doesn't care.


Maybe it's just in my circles like this, but I think these days chat itself is no service anymore. You need to integrate it into something else, like social networks, gtalk, skype (chat+telephone), "webcam-services". 10 years ago Grove might've been awesome, but now it feels like something is missing, because I can chat everywhere else contextrelated anyway.

Look how meebo developed over the last years. First it integrated web and traditional chat media like ICQ, MSN and then they already moved on to integrate their chat services into other peoples web-businessmodels as a b2b service.


> chat itself is no service anymore

And yet lots of shops pay 37signals for Campfire


Meebo also took $ SEVENTY million in funding.

When you take that sort of investment, the rules of the game change a lot I think. It's no longer sufficient to make a few million profit a month.


You are right, there is a possible niche that is probably profitable. Thus the businessmodel might be valid, even if they have no chance of hitting any main stream markets (with the businessmodel they have right now).


Nice work, Leah! Love the fact that it's just IRC at heart.

Is there any way to integrate a file-uploading facility (drag'n'drop in the web client, ideally, like Campfire) to the IRC concept, or is that an impedance mismatch?


Thanks!! I'd love to add some file uploading and/or hosting capability. I think it would be easy enough on the website side of things but I'm not sure yet what it would mean for existing IRC clients.


Let your services bot announce the user and url, perhaps?


Some sort of XDCC mechanism will be compatible with all modern clients.


Is it going to remain free? Would consider moving our internal irc server here but am concerned that in a few months it will become pay only?


No, it won't remain free forever. Soon we'll having pricing plans that are comparable to existing team chat products.


How unfortunate.

I've been working on an open source version of what you're purporting to provide as a service.


Open source software != free service -- in fact, even when the software is open source, you usually have to pay for a hosted, managed service.


I can make an automatic heroku application. I even have a guy who works at Heroku that would help me do it.

I could charge instance-hours for a prefabbed EC2 instance image too.


Do it.


Have a link for us?


"We are offering a free trial on all organizations during our beta testing period...... If you sign up for an organization we may contact you to gather feedback - both to improve Grove and help us determine our pricing."

I'd say that means it'll likely be a paid for service. Maybe with a small free option.


I've been beta testing irccloud.com. It's very well done, but I'd also like something that can be self-hosted to access internal-only IRC servers.


I like irccloud.com quite a lot. A bit different of a product, but still very nice.


Grove looks very cool. But groupme is free and has a good iphone/android client that a non-technical user can setup and use. Push notifications work great.

But Grove intrigues me for the ability to hook up useful bots to IRC.

If Groupme had an API available now, I'm not sure Grove would offer any benefits for my purposes. But I love how this space is exploding with all sorts of cool services.


Looks great, I'd love to try it, but I had an error 500 while attempting to register. Being an IRC fan, I believe a service like this will really help me to convince my workmates to talk over something else than Google Talk or Skype.


Just keep trying, I got 500s a couple of times too.


Sorry, our server is swamped right now. I'm working on it...

Update: I've made a few tweaks and it should be better now. fingers crossed


Works fine now! Thanks.


Cool idea. However, I am missing the method to auto-group chats by topic. Would be nice to have it with IRC.

Currently, I am doing that (not with IRC, but with Jabber/XMPP) via my side-project http://TwoToReal.com (beta): You ask questions (or raise topics) for which automatically determined experts are then pulled in via IM into a real-time web chat (most of the knobs/switches are tuned by machine learning).

What are your pricing plans?


Not sure whether it is in your vision as well, but wouldn't it be cool if Grove.io offered a library of such IRC agents in the same manner as HuBot from GitHub?


I think this is a super cool idea and it's something I've been thinking a lot about. Right now, existing IRC bots work fine with Grove but it would be cooler to have more integration.


Could even open a market of such bots for you: Customers of Grove.io could then offer their bots in your IRC bot market place ... but now I'm just dreaming ;-)


>Geeks love IRC, but it comes with a few hassles, mainly having to host it, that have led teams away from using it in favor of easier IM solutions.

The nature of IRC's channel-level organization¹ obviates the need to set up independent servers for a given team group.

What advantage does something like this offer me over just opening a new channel on freenode?

¹I.e., access lists (ops, voice) and modes like +psk (p[rivate], s[ecret], and k[eyed]).


Technically Freenode doesn't allow the creation of random channels for any purpose; they are required to be for some open project or open source software. (I say this as a member of quite a number of random channels on Freenode; but they could technically crack down at any time, and I would not recommend this approach for corporate use.)


But aren't the hassles one of the aspects us "geeks" like about IRC? I think us "geeks" deep inside love the terminal, having to setup an IRC client on a hosted server and tweaking the settings, getting our hands dirty. Us, "geeks", love also the sense of exclusivity that the barrier to entry to IRC provides. Take that, and the hassles, away and it becomes yet another chat network.


think you have a valid point there.


The Do team has been using Grove for a while now, and we're really liking it. The web app supports embedded content and gravatars, and connecting via IRC is great for those who love it.

We used Convore in the past, and Grove has definitely encapsulated the best parts of Convore along with laser-focused features and great performance. We actually can't wait to be a (or the first?) paying customer.


this is awesome, mixing a bit of irc + campfire. really like the ability for users to pick web interface or their own irc client.


Will its model work well with an open source/community project – where you want to allow all walk-up participants, but need some moderator powers?

For example, can you edit the searchable logs to remove spam?


I can actually think of a few good reasons why you might want to remove messages as a moderator. We'll add this soon.


And I already crashed my instance. Can no longer connect via IRC, only web client. Connected Hubot to it and did a help with all the scripts installed. Flooding ensued. Now it doesn't work.


what happened to convore?


I don't think people wanted to use it.


I dunno, I was pretty happy with it. It was really nice to have at conferences. Like any social medium, there's a tough chicken-and-egg nut to crack. I'm going to try to use it at Supercomputing next week.


Looks great, now time to convince my cofounders we need it...


I was thinking something like this would come in handy the other day. Goes to show if you have an idea, chances are someone else is already thinking about it.


Neat.

Easy hosted custom plugins/bots would be awesome.

You can run your own IRC bot, of course, but then you have to worry about finding a server to run it on, keep it running, etc.


Agreed. We need to figure out a good way to do better bot integration! (suggestions welcome)


I'd probably start with a configurable bot with a few predefined plugins (RSS feeds, version control / bug tracker integration, etc) and a configuration UI (similar to Github's Services hooks)

Eventually you could add a way to define custom plugins that are properly sandboxed.


You're describing http://hubot.github.com/ which ist pretty much a configurable bot with a few scripts and an irc adapter. It doesn't have a configuration page though.


I would suggest something like Twilio where you host your own xml at a specified url that twilio calls when it receives a call/sms for you. Similarly grove api consumers can specify a url that can be called by grove when a message is received in a channel and api consumers can respond ina predefined json.


I've been using this lightly for about a month now and it's a fantastic service. I hope it's wildly successful.


Now if it only were possible to get the github commit hooks working with this... Coding time I guess.


Too bad silcnet.org is dead. It would probably be a better base for such chat application than irc.


No /me implemented. :(


would be great to integrate with github


You should be able to use Hubot with Grove as it provides IRC interface. We have few users already using it.


Hubot does provide the irc-interface but then github itself doesn't provide a hubot hook. So you're pretty much stuck at the same place. Github itself does provide an IRC hook but that does not work since it can't authenticate with nickserv, it only allows for server and room passwords.


Assuming you'll have to create an entirely new Grove account for each bot you want to add, though?


That's fairly typical though.


i was wishing that this existed a few weeks ago. win. now i don't have to implement it.


Love the idea.


A website exists which is already here to pwn this space.

Facebook.


great idea


I've been using grove IRC for a while now. If you use the irc.grove.bz server instead of the .io server, you can create channels without registration.


oh wow this is amazing!! thanks!!




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

Search: