10 ms·
Show HN: Kimono – Never write a web scraper again
- toddwahnish 13y agoThis is fantastic. Congrats on launching it! Once it has pagination & auth I'll be all over this :)
- cbaleanu 13y agoDoes it do logging in to websites then fetching? Do you plan to add scripting to it?
- pranade 13y agoWe don't support logging in yet, but it's a feature we're working on adding. Scripting will also be cool, but it's right now further down our feature queue
- rmason 13y agoI thought to myself oh boy yet another web scraper as a service but got surprised. I haven't been this impressed with a product video since Dropbox.
- jmcgough 13y agoReally sleek interface, and looks like it could be extremely useful (I just spent a few hours cranking out Nokogiri this morning). Oh, typo: "Notice that toolbar at the toop of the screen?"
- pranade 13y agoAwesome, thanks for the kind words. And for catching that typo, will fix that now
- lstamour 13y agoI thought it was intentional. Swedish chef style. I like it, but I'll need to go back and re-read to understand how I can use this on other pages than the homepage/demo page. I've nothing immediate to try it with right now. :) Edit: I'll watch the video after work, probably will clear everything up for me.
- dikei 13y agoI don't think this can beat the speed of a hand-tune crawler. When I write crawlers, I skip rendering page and javascript execution if it isn't needed, which massively speed up the crawling process.
- pranade 13y agoThere's definitely things that a custom-built scraper can do more efficiently than kimono, but our focus right now making scraping accessible across a broad enough range of web sites.
- taternuts 13y agoThat looks quite swift
- tonystark 13y agoneat.
- ameister14 13y agoI really like how you guided me in to demoing. Nice job.
- rafeed 13y agoThis is awesome. Really nice implementation and so useful for many different applications. Just signed up and looking forward to trying this out.
- pranade 13y agoThanks guys, glad you like it. Welcome any feedback so we can make it better!
- sync 13y agoUndo button is awesome. More web apps need an undo button.
- szidev 13y agogreat idea. i'll have to keep this in mind for future projects.
- iurisilvio 13y agoWhat about some navigation tools there? Looks pretty good, but it does not really replace my scrappers. Maybe some of them...
- trip41 13y agore: nav tools -- you mean the ability to crawl multiple pages?
- dmunoz 13y agoI'm normally a bit worried when a thread quickly fills up with praise, but this looks very nice. It's something I have thought about, as I'm sure many people who have done any amount of scraping have, but never went forward and tried to implement. The landing page with video up top and in-line demo is a pretty slick presentation of the solution you came up with. Good job.
- pranade 13y agoThanks we were pretty surprised as well, but we're really grateful for the encouragement
- kenrikm 13y agoLooks awesome, however I keep getting errors and 404s. Could this be an issue on my end (seems to be working for others) or just HN making the servers beg for mercy?
- pranade 13y agoWhere are you getting the 404s? We will check into it now
- ForHackernews 13y agoThis looks really slick. What happens if a website you're scraping changes its design? Do you respect robots.txt?
- pranade 13y agoIf it changes the format significantly, the scraper will break, so for now you'll have to use the tool to rebuild. You will see on your API status page that it's down. As for robots.txt, we do respect it... for now we're leaving that to the user, but we're trying to implement a proactive way of checking for disallows and stopping those scrapers from being built.
- thatthatis 13y agoPlease clarify: are you saying that right now you leave respecting robots.txt to the user?
- pranade 13y agoAt the moment, we rely on users to be responsible. We spell it out in the terms and FAQ. We've been in private beta, keeping usage very limited until today. We fully understand the seriousness of the issue as we scale. We're committed to becoming a responsible bot that respects robots.txt
- loceng 13y agoI would say how you're scraping differs from say how Google, a search engine, scrapes. I'm not sure there is a way in robots.txt to define for each use? Knowing the data in a structured way, but then allowing it to be displayed in full off-site is quite different than using the scraped data for linking into a website.
- thatthatis 13y agoBut robots.txt provides minimums: don't scrape this page, don't refresh more than once every x, these crawlers are allowed this access, etc.
- jjcm 13y agoVery cool, and I like that the link is your announcement page running inside of the demo. Really drives home the idea. That said, it looks like it can't do media right now. I would love it if it could at least give me a url for images/other media.
- pranade 13y agoIt's a great suggestion, thanks! ... image extraction would be cool, and it's on our shortlist of features to build next
- twog 13y agoWell done on the product & solving a clear need! This is extremely useful for hackathons/prototyping. I also loved the live demo in the blog post and you did a wonderful job with the design/layout/colorscheme of the site.
- hcarvalhoalves 13y agoThis is excellent. Even it if doesn't work for scraping all sites, it simplifies the average use case so much that it's not even funny. Feature proposal: deal with pagination.
- pranade 13y agothanks. glad you like it. pagination, dynamic tabs (and crawling in general) is a big feature we really want to add soon. a lot of people are asking for it. the challenge will be integrating it with the current UEX which we're trying to keep super simple.
- kyriakos 13y ago+1 for paging. really important
- scotty79 13y agoAnother feature, simple one: Allow to add some filters to the data stream. For example: only posts that contain word "bitcoin" in the name or only those with 50 upvotes or more.
- pranade 13y agoThanks for the suggestion... adding to the list :)
- walden42 13y agoMake sure to include regex matching =)
- pranade 13y agoThanks. We support regex matching now. Try dragging to select text, if there's a relevant regex pattern and kimono will find it (there's an example inthe blog post). You can preview (and soon, you'll be able to edit) the CSS and Regex also in advanced mode.
- bluejellybean 13y agoHow (if at all) does this run on javascript heavy sites?
- pranade 13y agoJS-heavy sites can be tricky. We position it so it should execute after most of the on-page JS, so it handles a lot of cases. There are still sites that break it though.... we're trying to tackle these guys one by one right now, as we try to generalize a broader solution
- ph4 13y agoVery nice job. What about scraping data from password-protected pages?
- pranade 13y agoGreat request... it's on our feature shortlist. Definitely a feature we want to implement as soon as we can (after we tackle some basics like pagination and getting images)
- phillmv 13y agoThe UX is great and a journalists everywhere will thank you. But outside of government websites I don't see how a lot of this is even legal, per se?
- pranade 13y agoThanks... yes, public data from governments is a great use case. Often a lot of apps built using scrapers will wind up driving up traffic/ sales a the source site so it's okay. We want to do responsible web scraping, so will respect webmasters robots.txt files to make sure it's legal.
- hartard 13y agoI love the execution, but I also see inherent problems. Robots.txt is just a convention to advise crawlers. I'm confident most sites explicitly state this is against their terms of service. You will encounter terms along the lines of: "Unauthorized uses of the Site also include, without limitation, those listed below. You agree not to do any of the following, unless otherwise previously authorized by us in writing: Use any robot, spider, scraper, other automatic device, or manual process to monitor, copy, or keep a database copy of the content or any portion of the Site."
- pranade 13y agoYou've got a valid point. We want to eventually create a space that allows responsible scraping - so webmasters can have access to analytics on what's being scraped and can explicitly turn off kimono APIs for their domains if they see fit. We also think there are use cases for people who own their own data. Often, APIs will provide a way for companies to streamline their internal app development and figure out what to expose to the developer community before investing in an expensive API deployment.
- _delirium 13y agoThe law isn't entirely blind to conventions, though. They don't guarantee anything, but if a court understood that there exists a convention for saying "no robots, please", and the robot operator in question followed it, then a court could well look less favorably on the damages claims of a website operator who didn't make use of the widely known convention.
- georgemcbay 13y agoI've written more web scraping code than I care to admit. A lot of the apps that ran on chumby devices used scraping to get their data (usually(!) with the consent of the website being scraped) since the device wasn't capable of rendering html (it eventually did get a port of Qt/WebKit, but that was right before it died and it wasn't well integrated with the rest of the chumby app ecosystem). This service looks great, good work! But since you seem to host the APIs created how do you plan to get around the centralized access issues? Like on the chumby we had to do a lot of web scraping on the device itself (even though doing string processing operations needed for scraping required a lot of hoop jumping optimization to run well in ActionScript 2 on a slow ARMv5 chip with 64mb total RAM) to avoid all the requests coming from the same set of chumby-server IP addresses, because companies tend to notice lots of requests coming from the same server block really quick and will often rate limit the hell out of you, which could result in a situation where one heavy-usage scraper destroys access for every other client trying to scrape from that same source.
- pranade 13y agoAccess, legality and rate limiting issues come up a lot. We're working on a couple things to address them. The first is an intelligent job distribution system that consolidates scrapes across users and hits sites (and pages) at human-like intervals. the second is to create a portal for webmasters that allows them special privileged access to analytics on data being extracted from their sites, and the ability to "turn on or off" kimono APIs if they see fit. this way, via kimono, a webmaster at chumby could "provision" certain kimono users. we're still yet to see whether the later works out. thanks for the input
- tchadwick 13y agoThis looks really useful, and I'm trying to figure out if I could use it on a project I'm working on, but hitting an issue. I sent a support message. Nice job!
- pranade 13y agoThanks, the support tickets really help us debug. We'll look into it and get back to you
- trey_swann 13y agoThis is a great tool! In a past life we needed a web scraper to pull single game ticket prices from NBA, MLB, and NHL team pages (e.g. http://www.nba.com/warriors/tickets/single http://www.nba.com/warriors/tickets/single). We needed the data. But, when you factor in dynamic pricing and frequent page changes you are left with a real headache. I wish Kimono was around when we were working on that project. I love how you can actually use their "web scraper for anyone" on the blog post. Very cool!
- fnordfnordfnord 13y ago>Sorry, can't kimonify >According that web site's data protection policy, we were unable to kimonify that particular page. Sigh... Oh well... Back to scraping.
- pranade 13y agoWhat page were you trying to hit? We'll check it out
- fnordfnordfnord 13y agoPages buried in here: https://fannin4.wcjc.edu/ https://fannin4.wcjc.edu/ The course catalog is public, so no login is needed. I want to scrape various data related to courses, to populate forms automatically and such.
- nobodysfool 13y agoYeah, of course it won't work with HTTPS sites. They'd have to proxy those HTTPS sites and perform a MITM just to do it.
- Dylan16807 13y agoAs opposed to HTTP where they proxy it and MITM it? I don't understand the objection.
- fnordfnordfnord 13y agoI kind of just assumed that's what they were doing.
- garyjob 13y agoHTTPS is definitely a problem for proxy servers unless you the proxy server rewrites all the URLs in the html pages loaded as well as all the URLs of the Ajax calls to point back to the proxy server.
- guptaneil 13y agoNice work, this is much better than I expected! Does it require Chrome? It doesn't seem to work in Safari for me. Also, does Kimono work for scraping multiple pages or anything that requires authentication?
- pranade 13y agoGreat, it should work well on webkit browsers, what version of Safari are you using?
- guptaneil 13y ago7.0.1, the latest. I also don't have Flash installed, but it doesn't look like you're using Flash. The entire top bar doesn't show for me. Feel free to email me and I can send you screenshots.
- pranade 13y agoThanks - we're not using flash, so it must be something else. Will follow up over email
- cullenmacdonald 13y agothe reason i ever have to write a scraper is because of pagination. while this looks awesome, i'll have to stick to scraping until that is solved. :(
- pranade 13y agoIt's probably our #1 feature request at the moment. We're working on it and hope to have it ready for you to try soon
- randomdrake 13y agoThe presentation is beautiful and the website is great, but the tech broke so I have no idea how or if this even works. This is a wonderful concept and one I've talked about doing with others. I was really excited to try this. I watched the demo video and it seemed straightforward. I went to try and use it on the demo page it provides, going through and adding things, but when I went to save it, I just received an error that something went wrong. Well, crap. That was a waste of time. Oh well, maybe it's just me. Alright, I'll give it another shot using the website they used in the demo. Opened up a Hacker News discussion page and started to give it a try. Immediately it was far less intelligent than the demo. Clicking on a title proceeded to select basically every link on the page. Somehow I clicked on some empty spots as well. Nothing was being intelligently selected like it was in the demo. Fine, that wasn't working tremendously well, but I wanted to at least see the final result. Same thing: just got an error that something went wrong and it couldn't save my work. Disappointing. I still might try it again when it works 'cause it's a great idea if they really pulled it off. So far: doesn't seem to be the case.
- pranade 13y agoSorry you had a bad first experience. We've tested this on a lot of sites, and it works stably across a lot of different cases, but haven't solved it everywhere. Thanks for letting us know about the discussion page. We'll look into the bugs right now
- randomdrake 13y agoHaving written a whole lot of crawling code throughout the years, I can totally understand how monumental of a task this is. This really does look like a cool product. Glad you're actively hunting down ways to improve the demo before asking for money :) Feel free to drop me a line if you want any specifics about the troubles I had.
- faceyspacey 13y agohacker news has probably been scraped too many times. that's why it's not working. it didn't work for me, but other sites worked easily. ...i will comment that the user experience for first adding elements is not what i expected. once u understand it's great overall, but i'd have some popup tooltip type guide thing going on for the first time a user uses it explaining that you should choose one element first, and then check a matching element. ...Then explain the number bubbles next to the property field. You don't get what those numbers are at first. ..Then for the property field, somehow make it more obvious that you should fill it out. Basically you have the interface setup so any action can be done at any time, but it should be presented as a forced sequential set of steps to the end user, at least at first.
- aqme28 13y agoI would seriously consider rethinking that Favicon.
- misuba 13y agoSeconded. I can't show that to anyone at work.
- roryokane 13y agoTo me, the favicon merely looks like a sumo wrestler’s head with a short ponytail and scowling/serious eyebrows. I can’t tell what NSFW thing you see it as.
- hobs 13y agoSpitballing here, but it could be a POV dick's eye view. I am really reaching however. edit: Interesting downvotes, I dont mean to be rude, I am just trying to see what someone could find offensive. (I dont see the problem with the favicon)
- prolways 13y agoUntil I read this thread I also saw a sumo or an angry onion, but I believe the picture is actually a person facing away from us undoing their kimono.
- rfnslyr 13y agoWhy? It looks like an onion with angry eyebrows. Do you have something against vegetables?
- xux 13y agoWow looks amazing. I tried doing some queries on public directories, and it even supports parameter passing. Will be using this for some side projects.
- jfoster 13y agoCool concept. One concern I'd have about this type of tool is that when it encounters something it can't handle, I'm stuck. Writing your own scraper means that you can modify it when you need to. I think the ultimate solution would be something like Kimono with the ability to write snippets of custom javascript to pull out anything that it can't handle by default.
- pranade 13y agoWe're in the middle of implementing a more power developer version of the tool to handle the use cases you're talking about. The beginning of this is surfaced under the "advanced" tab in the data model view where we show the selectors and regular expressions that are produced. we want to ultimately let you edit those to customize the extractor. From there it'd be super cool to implement the javascript snippet feature you suggested.
- nnnn 13y ago"Never write a web scraper again"... yea right.. sick and tired of such gimmicks and self promotion on the net today.
- chevreuil 13y agoWe all know there are a lot of existing tools that does the same things. But I've not met one with such a polished UX. Kudos to the Kimono team, I'll definitly recommend your product.
- tlrobinson 13y agoI built something very similar last year, but sadly never got around to polishing and launching it: http://exfiltrate.org/ http://exfiltrate.org/ (There's a prototype of an API generator hidden in a menu somewhere but it's nowhere near production ready)
- pranade 13y agoYeah, we've been working on this for a while too... took a while to polish it a bit before we could put it out there. Will check out exfiltrate.org - looks cool!
- rlpb 13y agoAre you familiar with ScraperWiki? I'm wondering how your work fits in with it. Edit: looks like they've moved away from that space, but have an old version available at: https://classic.scraperwiki.com/ https://classic.scraperwiki.com/
- Maxious 13y agoThe people who scrape data to avoid paying for APIs are the same people who will not pay for a service to make scraping easier ;)
- handelaar 13y agoThe people who scrape data at Scraperwiki -- which was made by the same people who opened up parliamentary transcripts in the UK for the first time, and the UN's proceedings, and data about how MPs in London vote -- generally don't have an option to buy anything because the data's hidden by governments from the people who paid for it, on purpose. But by all means take this opportunity to dismiss all of us as freeloaders.
- AznHisoka 13y agoActually, I usually scrape data because there is NO API I can pay for.
- ThomPete 13y agoThank you for building a tool I been wanting so I don't have to! Can't wait to play around with this tonight. Suggestion. Allow one to select images.
- dunham 13y agoThe Simile group at MIT did something similar back around 2006. Automatic identification of collections in web pages (repeated structures), detection of fields by doing tree comparisons between the repeated structures, and fetching of subsequent pages. The software is abandoned, but their algorithms are described in a paper: http://people.csail.mit.edu/dfhuynh/research/papers/uist2006-augmenting-web-sites.pdf
- pranade 13y agoThanks a ton for sharing... the association algorithms have been where we've been spending a good chunk of time. Will read through this
- tlianza 13y agoIf you're interested in hosted solutions that try to do automatic identification of pages, diffbot is worth a look. We've had some good experiences: http://diffbot.com/ http://diffbot.com/
- losvedir 13y agoOh, hey, memories. I worked one summer with David Huynh (who you're linking to there) and David Karger (his thesis advisor) on one of the Simile projects. I vaguely remember playing around with this tool you mentioned. I thiiiiink it was this one[0], although it seems to be superseded by this one[1] now. [0] http://simile.mit.edu/wiki/Piggy_Bank http://simile.mit.edu/wiki/Piggy_Bank [1] http://simile.mit.edu/wiki/Sifter http://simile.mit.edu/wiki/Sifter
- danso 13y agoJust had to chime in and say that David Huynh and his fellow programmers will be forever heroes to me and a small group of data journalists who depended on Gridworks/Google Refine/OpenRefine
- dome82 13y agoI like the concept and it looks similar at Import.io
- pranade 13y agoYes, in concept quite similar. We wanted to make something that you can use from within your browser as part of your natural workflow, without installing any other software.We also really wanted to figure out the right data association intelligently based on user selections vs. asking users to think through a data model up front
- fsckin 13y agoConstructive Tone: I figured that it might be nifty to scrape cedar pollen count information from a calendar and then shoot myself an email when it was higher than 100 gr/m3. This would be a pretty difficult thing to grab when scraping normally, but the app errors before loading the content: https://www.keepandshare.com/calendar/show_month.php?i=1940971 https://www.keepandshare.com/calendar/show_month.php?i=19409... JS error: An error occurred while accessing the server, please try againError Reference: 6864046a
- pranade 13y agoThanks for letting us know. Just tried and am getting the same error. The page is loading content dynamically from another source... We'll look into this and see if we can get it working on this page
- fsckin 13y agoDo you support POSTs for fetching dynamic data? I found where it's pulling from, here's the curl command: curl "https://www.keepandshare.com/calendar/fns_asynch_api.php?r=0.04834879608824849&fromapp=calendar" https://www.keepandshare.com/calendar/fns_asynch_api.php?r=0... --data "action=getrange&i=1940971&from=2013-12-26&to=2014-02-06"
- pranade 13y agoNo, we don't have POST support quite yet. We're working on a solution.
- catshirt 13y agoreally excited to see this. i've had the idea (and nearly this execution) in mind for years but no use or ambition to get it done. given the pricing though i'm almost motivated to make my own. as a hosted service the fees make sense with the offerings. but not only would i rather host my own- it would be cheaper all around. would you consider adding a free or cheap self hosted option? aside, i think there is a mislabel on the pricing page. i'm guessing the free plan should not have 3 times the "apis" than the lite plan.
- wprl 13y agoIt's easy not to write web scrapers even without this tool ;)
- shekyboy 13y agoLike the parameter passthrough feature. Take a look at places where the parameters are part of the URL structure. For example a Target product page http://www.target.com/p/men-s-c9-by-champion-impact-athletic-shoe-black/-/A-14656388#prodSlot=medium_1_1 http://www.target.com/p/men-s-c9-by-champion-impact-athletic... In order to get data for a different product, I will have to modify the URL itself. I think same holds true for blog posts.
- pranade 13y agoYes, it's a great point. We're working on updating the query param passthrough to handle params within the URL structure.
- shekyboy 13y agoApart from that here are other items. You may have these on your list, but can count my vote to prioritize. 1. Pagination 2. Image URLs 3. Focus on page types such as product pages, posts etc. That way its easy to go from content to content. Will help crawling too 4. Link back to original page included in JSON Finally common sites/pages used by multiple users of your systems should not count against the API count requirement under pricing. You may want to charge against total calls, like Parse.
- pranade 13y agoThanks for the suggestions!
- DanBlake 13y agoShow me it working with authentication and you will have a customer. Scraping is always something you need to write because the shit you want to get is only shown when you are logged in.
- pranade 13y agoYes, it's one of the most popular feature requests. We don't support auth yet, but it's on our shortlist and we hope to have it ready soon.
- chii 13y agohow are you going to do it without having to know the actual authentication key(s)? if i don't trust anyone enough to give my auth away, and so unless the site being scraped has some sort of oauth support, how are you going to get any data? of course, if this was an offline product, or self-hosted product, then it would solve that problem of auth instantly.
- the_french 13y agoWould there any way to fake the beginning of an OAuth session with Facebook, Google or any other OAuth authenticated site? Kind of like replaying cookies to hijack sessions?
- garyjob 13y agoThe route of proxying the web page presents much difficulty in doing actual authentication on Facebook or Google's website via the proxied webpage without first rewriting most of the javascript and hijacking their Ajax calls on the fly. The approach I took was to hijack the Cookies from the browser once the user has signed in after on e.g. Facebook via the browser extension. The route of proxying the website does in fact do away with the need to install any external 3rd libraries. This browser extension I built coupled with the web service its integrated to does allow for scraping of pages from Facebook, Google and LinkedIn logged in pages as well. https://chrome.google.com/webstore/detail/krakeio/ofncgcgajhgnbkbmkdhbgkoopfbemhfj https://chrome.google.com/webstore/detail/krakeio/ofncgcgajh...
- pknight 13y agoThat UI made me go wow, this could be an awesome tool. Idea that pops into my mind is being able to grab data from those basic local sites run by councils, local news papers etc and putting it into a useful app. How dedicated are you guys to making this work because I'd imagine there are quite a few technical hurdles in keeping a service like this working long term while not getting blocked by various sites?
- pranade 13y agoLove your suggestion. We're committed to making kimono better and we're working on it all the time. We want to make sure it's a responsible scraper, so want to work together with webmasters in cases where there might be blocking but the data is legal to share...
- dmritard96 13y agoas someone building a home grown proprietary scraping engine. Consider alternative locations of elements. Most sites are using templating engines so its fairly reliable to find things in the same place, but more often than you might expect, things move a round ever so slightly. Navigation is a fun one also. ;)
- NicoJuicy 13y agoThis is really slick! Btw. Who made your intro video?
- pranade 13y agoIt's pretty homegrown right now, we did the intro video ourselves on our laptops
- ewebbuddy 13y agoReally cool idea and tool. Still need to test this out properly. Is it possible to scrape note just one page but a stack of them? For example - a product catalog of 1000 SKUs extending upto 50pages.
- pranade 13y agoWe don't support that quite yet. It's our #1 feature request though, and we're working to get it ready soon
- iamkoby 13y agoi love this! and amazing video!
- blazespin 13y agoThere's a huge business here if you keep at it. I'll throw money at the screen if you can make this work.
- harryovers 13y agoso what do you do that import.io doesn't?
- alternize 13y agolooks promising! to be fully usable for me, there are some features missing: - it lacks manual editing/correcting possibilities: i've tried to create an api for http://akas.imdb.com/calendar/?region=us http://akas.imdb.com/calendar/?region=us with "date", "movie", "year". unfortunately, it failed to group the date (title) with the movies (list entries) but rather created two separate, unrelated collections (one for the dates, one for the movies). - it lacks the ability to edit an api, the recommended way is to delete and recreate. small bugreport: there was a problem saving the api, or at least i was told saving failed - it nevertheless seems to be stored stored in my account
- pranade 13y agoThanks for the feedback. We're working on a feature that will allow you to edit APIs you've created and also edit the selectors and regex (right now, in advanced mode, you can see them, but cannot edit). We're looking into your bug now...
- jlees 13y agoI like how you've thought through the end to end use case: not just generating an API, but actually making it usable. I've done my fair share of web scraping and it's not an easy task to make accessible and reliable -- good luck! It makes me wonder if there isn't a whole "API to web/mobile app with custom metadata" product in there somewhere. I can imagine a lot of folks starting to get into data analysis and pipelines having an easier time of it if they could just create a visual frontend in a few clicks.
- pranade 13y agoYes, we're excited about the possibilities of an end-to-end use case as well... in fact we were surprised when we found a lot of interest in front-end output layers on top of the APIs than just the APIs themselves. Would be curious to know what output features would be most valuable for you.
- jlees 13y agoWell. Think about spreadsheets. Think about live spreadsheets powered directly by APIs. Bingo! I'm doing a couple of data mining projects right now and simply being able to query and look into the API outputs, as well as my local database, without building a custom frontend would've saved me a bunch of time. But I'm thinking more of the knowledge worker, or even the power user who wants to view their Fitbit, Up and Lark data all in the same dashboard. Can't help but thinking this already exists somewhere though.
- pranade 13y agoLove your idea. Would love to follow up on this with you
- tectonic 13y agoJust use http://selectorgadget.com http://selectorgadget.com
- aaronsnoswell 13y agoMan that demo is impressive!
- dclara 13y agoAgree demo is awesome. But I don't think scraping for any web page is that simple. Lots of exceptional cases.
- BinaryBird 13y agoNice tool, slick UI. It worked for some pages and not for others. Currently I'm using Feedity: http://feedity.com http://feedity.com for all business-centric data extraction and it has been working great (although not as flexible as kimono).
- pyed 13y agoactually I love scraping :(
- abvdasker 13y agoI kind-of enjoy writing web scrapers.
- bigd 13y agoSeems it can't see the stuff inside angular views.. well at least mines.. But for the rest, awesome product. Thanks.
- pranade 13y agoYes, you're right... we can't handle angular quite yet. We're working on this.
- jval 13y agoGreat job guys. One problem I've had though is that I think you guys are hosted on AWS - a lot of websites block incoming connections from AWS. Are there plans to add an option in future to route through clean IPs? Premium or default, this would be cool and make it a lot more useful.
- eth 13y agoGreat tool! I'm coming at things from a non-coder perspective and found it easy to use, and easy to export the data I collected into a usable format. For my own enjoyment, I like to track and analyze Kickstarter project statistics. Options up until now have been either labor intensive (manually entering data into spreadsheets) or tech heavy (JSON queries, KickScraper, etc. pull too much data and my lack of coding expertise prevents me from paring it down/making it useful quickly and automagically) as Kickstarter lacks a public API. Sure, it is possible to access their internal API or I could use KickScraper, but did I mention the thing about how I dont, as many of you say, "code"? What I do understand is auto-updating.CSV files, and that's what I can get from Kimono. Looking forward to continued testing/messing about with Kimono!
- rismay 13y agoOMFG.
- rpedela 13y agoDefinitely awesome presentation and product. The example doesn't seem to work right on Firefox. On Chrome, if I click "Character" in the table then it highlights the whole column and asks if I want to add the data in the column. On Firefox, clicking "Character" just highlights "Chatacter" and that is it. Ubuntu 12.04 Firefox 25.0.1
- pranade 13y agoThanks for flagging. We'll get on this to figure out what's going on when running on ubuntu
- mswen 13y agoHow does this compare with Mozenda?
- thatthatis 13y agoThis is my third time trying to get an answer to this question: does your crawler automatically respect robots.txt?
- pranade 13y agoSo sorry for missing this earlier. See our response in comments below: "At the moment, we rely on users to be responsible. We spell it out in the terms and FAQ. We've been in private beta, keeping usage very limited until today. We fully understand the seriousness of the issue as we scale. We're committed to becoming a responsible bot that respects robots.txt"
- thinkzig 13y agoGreat work so far. The tool was very intuitive and easy to use. My suggestion: once I've defined an API, let me apply it to multiple targets that I supply to you programatically. The use case driving my suggestion: I'm an affiliate for a given eCommerce site. As an affiliate, I get a data feed of items available for sale on the site, but the feed only contains a limited amount of information. I'd like to make the data on my affiliate page richer with extra data that I scrape from a given product page that I get from the feed. In this case, the page layout for all the various products for sale is exactly the same, but there are thousands of products. So I'd like to be able to define my Kimono API once - lets call it CompanyX.com Product Page API - then use the feed from my affiliate partner to generate a list of target URLs that I feed to Kimono. Bonus points: the list of products changes all the time. New products are added, some go away, etc. I'd need to be able to add/remove target URLs from my Kimono API individually as well as adding them in bulk. Thanks for listening. Great work, again. I can't wait to see where you go with this. Cheers!
- pranade 13y agoThanks a ton for the feedback. Getting data from multiple similarly structured URLs programmatically is something we're working on now. We love hearing about the use cases you want to use this for so we can make sure we build out the right features to make kimono useful for you.
- joshmlewis 13y agoIs there an ability to scrape more than one page of data?
- pranade 13y agoNot yet, but it's our #1 feature request, so we're working on it now. For now, you can make multiple APIs (one for each URL). If the URL takes query parameters though, you can re-use the same API and programmatically cycle through query parameters.
- GigabyteCoin 13y agoI'm curious how you plan to avoid/circumvent the inevitable hard IP ban that the largest (and most sought after targets) will place on you and your services once you begin to take off? I could have really used a service like this just yesterday actually, I ended up fiddling around with iMacros and got about 80% of what I was trying to achieve.
- pranade 13y agoIt's a great question. What we're really trying to do is make data accessible programmatically and at scale. We want to connect data providers and data consumers with APIs in a way that's mutually beneficial vs. being a tool for data theft. Our hope is to (once we scale) actually work with data providers directly on the on the distribution of their data so the IP ban becomes a non-issue.
- Nilzor 13y agoBut isn't the point kinda to let the users come up with data providers themselves? If you say "Only these 500 data providers are available for scraping", you don't have a business. If you don't' have such a limitation, you'll not be able to work directly with all data providers. You'll have IP problems.
- garyjob 13y agoI found the one click action for selecting an entire column of values as well as the UI/UX on the top column of the page to be very impressive. We were thinking of a nice clean way to represent that particular UI/UX flow in this browser extension we built as well. Will incorporate that in our next release. https://chrome.google.com/webstore/detail/krakeio/ofncgcgajhgnbkbmkdhbgkoopfbemhfj https://chrome.google.com/webstore/detail/krakeio/ofncgcgajh... Would love to meetup and exchange some ideas if you are based in Bay area.
- keyurfaldu 13y agoAwesome! Hats off.. How about extracting hashtag/GID of any record if applicable, which are typically not rendered on page, but hidden under the hood.
- IbJacked 13y agoWow, this is looking good, I wish I had it available to me 6 months ago! Nice job :D I don't know if it's just me or not, but it's not working for me in Firefox (OSX Mavericks 10.9.1 and Firefox v26). The X's and checkmarks aren't showing up next to the highlighted selections. Works fine in Safari.
- pranade 13y agoThanks for letting us know. We've tested on some versions of Firefox, but not v26 on Mavericks. We'll look into this
- yummybear 13y agoLooks very nice. There seems to be an issue with international characters though (æ/ø/å).
- pranade 13y agoYes, thanks for spotting. We'd discovered that Chinese, Japanese and Korean failed but didn't know about these characters. Thanks!
- bambax 13y ago> Web scraping. It's something we all love to hate. You wish the data you needed to power your app, model or visualization was available via API. But, most of the time it's not. So, you decide to build a web scraper. You write a ton of code, employ a laundry list of libraries and techniques, all for something that's by definition unstable, has to be hosted somewhere, and needs to be maintained over time. I disagree. Web scraping is mostly fun. You don't need "a ton of code" and "a laundry list of libraries", just something like Beautiful Soup and maybe XSLT. The end of the statement is truer: it's not really a problem that your web scraper will have to be hosted somewhere, since the thing you're using it for also has to be hosted somewhere, but yes, it needs to be maintained and it will break if the source changes. But I don't see how this solution could ever be able to automatically evolve with the source, without the original developer doing anything?
- littledot5566 13y agoPerhaps this could be automated by finding the same content in two versions of the dom and then doing a diff on the structure, updating the rules?
- pranade 13y agoIt would be great to automate this eventually. For now, we're trying to make it really easy to set up and rebuild the scraper. If it goes down, you'll see it in the status on your user dashboard. We're also implementing alerts, so you can opt to get an email notification if a scrape fails
- rpedela 13y agoI assume you get an error on the hourly, daily, monthly, whatever update which you are notified about. Then you can redo the semi-manual setup of the scraper.
- diegolo 13y agoIt would be nice to have a view also on the raw html code, e.g., to create a field containing the url of an image in the page.
- pranade 13y agoThanks, for the suggestion. We're rolling out advanced mode soon, which will allow you to edit the CSS selectors and RegEx operating on the page's HTML to define the selected data elements
- lips 13y agoI'm experiencing login errors (PEBKAC caveat: password manager, 2x checked, reset), but the support confirmation page is a nice surprise. http://i.imgur.com/w01CoUy.jpg http://i.imgur.com/w01CoUy.jpg
- PhilipA 13y agoIt looks cool, but very expansive compared to Visual Web Ripper, which you pay way less for (but has to host yourself).
- byteface 13y agouse any chrome xpath plugin and give that to YQL
- kyriakos 13y agoIt appears that it doesn't work with websites containing international characters.
- pranade 13y agoYes, thanks for noting. This is a bug and we're working to get these characters supported as soon as we can
- paul1664 13y agoReminds me of Dapper http://open.dapper.net/ http://open.dapper.net/ This allowed you to do similar, before being consumed by Yahoo. Might be worth a look.
- fosfori 13y agoJust want to point out that you can use the tool without logging in if you block the prompt with e.g. AdBlock. And yeah, it seems to work just fine in the HN front page.
- lucasnemeth 13y agoNice job! I really liked, it's a fantastic idea! And your UX is great! Just one thing I've found when testing: I've had some problems with non-ascii characters, when I was visiting brazilian websites, such as this : www.folha.com.br.
- timov 13y agoYou can use the utility without registration or login by blocking the login prompt with, for example, AdBlock.
- narzero 13y agoI like the concept. Would love to see page authentication
- shamsulbuddy 13y agoDoes such Webscraping is allowed legally. Since it is not done directly from our servers and if any legal action will be taken by the scraped website , will it be on kimonolabs..or the user..
- critium 13y agoPlease get this off the ground. I would also possibly suggest a separate business, website regression testing. Selenium is WAAAY to painful.
- pranade 13y agoThanks for the suggestion... we're working hard to get auth up and running!
- mhluongo 13y agoAny chance you guys plan to add link hrefs to CSVs? I'd love to use this now, but I need the href for backlinks and future inference.
- pranade 13y agoThanks for the suggestion, we're adding to our list
- cycnusx 13y agoOne of my favorites: http://htmlagilitypack.codeplex.com http://htmlagilitypack.codeplex.com