Good alternatives to Xenu's Link Sleuth and AuditMyPc.com Sitemap Generator
-
I am working on scraping title tags from websites with 1-5 million pages. Xenu's Link Sleuth seems to be the best option for this, at this point. Sitemap Generator from AuditMyPc.com seems to be working too, but it starts handing up, when a sitemap file, the tools is working on,becomes too large. So basically, the second one looks like it wont be good for websites of this size. I know that Scrapebox can scrape title tags from list of url, but this is not needed, since this comes with both of the above mentioned tools.
I know about DeepCrawl.com also, but this one is paid, and it would be very expensive with this amount of pages and websites too (5 million ulrs is $1750 per month, I could get a better deal on multiple websites, but this obvioulsy does not make sense to me, it needs to be free, more or less). Seo Spider from Screaming Frog is not good for large websites.
So, in general, what is the best way to work on something like this, also time efficient. Are there any other options for this?
Thanks.
-
import.io and it's free
-
Another idea that I have here, is to look for sitemaps of these websites. There may be a way to get a list of all the urls, right away, without crawling. Look at /robots.txt, /sitemap.xml, search for sitemap in Google, things like that. If there is urls, title tags can be scraped with Scrapebox, and as far as their website is saying, it can be done relatively fast.
# # Edit:
I had somebody suggesting http://inspyder.com, around $40 and free trial. May be a good option too.
-
So there is probably no way to tell, whether I have all the urls of the site, or what percentage I have... I may have 80 or even less percent of the total site, and not know about it, I would assume. This is one of the parts of working on the sites (I've never needed it, but I am working on something like this now), and there is no good tool, which would do the work.
I have a website with 33,500,000 pages. I've been running the tool for close to 5 hours, and I have around 125,000 urls, so far. This means, that it would take 1340 hours to do the entire site. This is close to two months of running the program 24 hours a day, which does not make sense. And besides that I was planning to do it on up to 100 sites. Definitely not something that can be done, and I would say that it should be possible, software-wise.
I will try your method, and see what I will get. I dont have too much time for experimenting with it too. I need to work, and generate results...
# # Edit
I will now how the number of urls compares to the 33,500,000 figure, obviously, but whats indexed in Google is not necessarily the complete website too. The method that you are suggesting is not perfect, but I dont have two months to wait too, obviously...
-
You will crawl some of the same URLs - that's why you remove duplicates at the end. There's no way to keep it from re-crawling some of the URLs, as far as I know.
But yes, get it to recognize 600-800k URLs and then split the file. (Export, put the links in as an html file and start over.) Let me break it down the best I can:
-
Crawl your main (seed) URL until you've recognized 800k.
-
Pause/stop and then export the results.
-
Create an html file with the URLs from the export - separated 50k to 100k at a time.
-
Recrawl those files in Xenu with the "file" option.
-
Build them back up to 800k or so recognized URLs again and repeat.
After a few (4-6) iterations of this, you'll have most URLs crawled on most sites no matter how large. Doing it this way, I think you could expect to crawl about 2-3 million URLs a day. If you really paid attention to it and created smaller files but ran them more frequently, you could get 4-5 million, I think. I've crawled close to that in a day for a scrape once.
-
-
Thanks. It is good to hear, that there is a way to do, of what I am trying to do, especially on 50 or more sites, large.
I've been running Xenu on a 33,500,000 pages site for a little over 4 hours and 15 minutes, and I have something like this, so far:
Close to 500,000 urls recognized, and only 115,000 processed, it looks like. I am manually saving it to a file, every now and then, as there is no way to auto save, as far as I was checking (there could be though, I am not sure, there is no too many options there).
I am not sure, based on your advice, how I could speed it up this process. Should I wait from this point, then stop the program, and divide the file into 8 separate files, and load it to the program separately? Then the program will recognize these separate files as one, and it will continue crawling for new urls? If possible, please give better information on how this would need to be done, as I dont fully understand. I also dont see how this could do this large website in one day, or lets say even five days...
# # Edit:
I actually got to understanding what you mean, get 8 separate files (can be 6 or, lets say 10) and run them all at the same time. But still, how will the program know not to crawl and download the same urls, on all the files? In general, I would like to ask for better explanation, on how this needs to be done.
Thanks.
-
Let Xenu crawl until you have about 800k links. Then export the file and add it back as 8 x 100k lists of URLs. You can then run it again and repeat the process. By the time you have split it 4-5 times, you can then export everything, put it into one file and remove duplicates.
Xenu, done this way, with 100 threads, is probably the fastest way to do the whole thing. I think you could get the 5M results in under 1 day of work this way.
-
Ok. So it looks like Screaming Frog may be a good way to go too, if not better. Xenu is free, which is a big plus. On the top of that Creaming Frog's Seo Spider is based on a yearly subscription, and not a one time fee. For those who dont know, there is a version of Xenu for large sites, which can be found on their website. They also have a support group at groups.yahoo.com (find it through there), I am not sure if it is still active.
Xenu upgraded to the version for larger sites may be the best way to go, since it is free. I've been testing AuditMyPc.com Sitemap Creator and the better version of Xenu, and the first one already hanged up (I discontinued using it). They were both collecting the info at about the same speed, but Xenu is working better (does not hang up, looks like it should be good). Either way, this will take quite a lot of time, with it, as previously mentioned.
-
I agree with Moosa and Danny - in terms of I use Screaming Frog (full paid version) on a stripped down windows machine with an SSD and 16GB of performance RAM. I have also download the 64 bit version of Java and increased the memory allocation for Screaming Frog to 12GB (default limit is 512mb) - here's how - http://www.screamingfrog.co.uk/seo-spider/user-guide/general/ (look at the section Increasing Memory on Windows 32 & 64-bit)
I did this as I was having issues crawling a large site - after I put this system in place it eats any site I have thrown at it so far so it works well for me personally. In terms of speed of crawl large sites such as you mention will still take a while - you can set crawl speed in Screaming Frog, but you need to be careful as you can overload the server of the site you are crawling and cause issues...
Another option would be to buy a server and configure it for Screaming Frog and other tools you may use - this gives you options to grow the system as your needs grow. It all depends on budget and how often you crawl large sites - obviously buying a server such as a windows instance on Amazon EC2 will cost more in the long run but it takes the strain away from your own systems and networks plus you should effectively never hit capacity on the server as you can just upgrade. It will also allow you to remote desktop in on whatever system you use - yes even a Mac
Hope this helps
-
I believe when you are talking about 1 to 5 million URLs it is going to take time no matter what tool you use but if you ask me screaming frog is a better tool and if you have a paid version of it you still can crawl websites with few million URLs in it.
Xenu is not a bad choice either but it’s kind of confusing and there is a possibility that it can broke.
Hope this helps!
-
I was facing similar issue with huge sites, that have over 100s of thousands of pages. But ever since I upgraded my computer with RAM and SSD it run way better on huge sites as well. I tried several scrappers and I still believe Xenu is the best one and most recommended by SEO experts. Also you might want to check this post on Moz Blog about Xenu's
http://moz.com/blog/xenu-link-sleuth-more-than-just-a-broken-links-finderGood luck!
Got a burning SEO question?
Subscribe to Moz Pro to gain full access to Q&A, answer questions, and ask your own.
Browse Questions
Explore more categories
-
Moz Tools
Chat with the community about the Moz tools.
-
SEO Tactics
Discuss the SEO process with fellow marketers
-
Community
Discuss industry events, jobs, and news!
-
Digital Marketing
Chat about tactics outside of SEO
-
Research & Trends
Dive into research and trends in the search industry.
-
Support
Connect on product support and feature requests.
Related Questions
-
Could using our homepage Google +1's site wide harm our website?
Hello Moz! We currently have the number of Google +1's for our homepage displaying on all pages of our website. Could this be viewed as black hat/manipulative by Google, and result in harming our website? Thanks in advance!
Technical SEO | | TheDude0 -
Why is Google's cache preview showing different version of webpage (i.e. not displaying content)
My URL is: http://www.fslocal.comRecently, we discovered Google's cached snapshots of our business listings look different from what's displayed to users. The main issue? Our content isn't displayed in cached results (although while the content isn't visible on the front-end of cached pages, the text can be found when you view the page source of that cached result).These listings are structured so everything is coded and contained within 1 page (e.g. http://www.fslocal.com/toronto/auto-vault-canada/). But even though the URL stays the same, we've created separate "pages" of content (e.g. "About," "Additional Info," "Contact," etc.) for each listing, and only 1 "page" of content will ever be displayed to the user at a time. This is controlled by JavaScript and using display:none in CSS. Why do our cached results look different? Why would our content not show up in Google's cache preview, even though the text can be found in the page source? Does it have to do with the way we're using display:none? Are there negative SEO effects with regards to how we're using it (i.e. we're employing it strictly for aesthetics, but is it possible Google thinks we're trying to hide text)? Google's Technical Guidelines recommends against using "fancy features such as JavaScript, cookies, session IDs, frames, DHTML, or Flash." If we were to separate those business listing "pages" into actual separate URLs (e.g. http://www.fslocal.com/toronto/auto-vault-canada/contact/ would be the "Contact" page), and employ static HTML code instead of complicated JavaScript, would that solve the problem? Any insight would be greatly appreciated.Thanks!
Technical SEO | | fslocal0 -
Are the duplicate content and 302 redirects errors negatively affecting ranking in my client's OS Commerce site?
I am working on an OS Commerce site and struggling to get it to rank even for the domain name. Moz is showing a huge number of 302 redirects and duplicate content issues but the web developer claims they can not fix those because ‘that is how the software in which your website is created works’. Have you any experience of OS Commerce? Is it the 302 redirects and duplicate content errors negatively affecting the ranking?
Technical SEO | | Web-Incite0 -
What's the rules on overly dynamic URLs ?
Developer says "Overly-Dynamic URL. Developer says that this is the hardest and complex part. It will be possible to change all of the search criterias to use ( / )
Technical SEO | | stewbuch1872
But in this case each of the pages will be indexed and every time listing gets added, content will get changed. Which for example Google will start blocking what is the best way to address this and will google block as suggested ? thanks0 -
Wordpress New Category URL's
Were just about to redesign our site and put all the blogs over to the new site. Previously most blogs have been added to the uncategorised section of the blog and I'm now weighing up the benefit of sifting through all the blogs and changing them to relevant categories. From an SEO perspective would it be better to Leave them in their current category but start afresh with all new blogs by adding them to relevant categories? Work out which blogs should go in which new category and 301 all previous URL's to the new one. Obviously number one will take a lot more time than number two.
Technical SEO | | acs1110 -
On-Page Report Says 'F', and I'm Confoozled As to Why
I'm primarily interested in how we failed in our "Broad Keyword Usage in Title" category. The Keyword Pair we're gunnin' for is: "Mac Windows" Our current page title is: "CrossOver: Windows on Mac and Linux with the easiest and most affordable emulator - CodeWeavers" This is, I grant, ugly. However, bear with me. SEOMoz Report Card says "Easy Fix!" and suggests: "Employ the keyword in the page title, preferrably as the first words in the element." I humbly submit that "Mac" and "Windows" IS in the page title. So what am I missing? Is it the placement of the words relative to each other, or relative to the start of the sentence? Or is the phrase "CrossOver:" somehow blocking the rest of the sentence from being read? Are colons evil? I'm genuinely mystified as to why (from a structural standpoint) our existing title tag is failing this test, and I'd be delighted for answers and/or feedback. Thanks in advance.
Technical SEO | | CodeWeavers0 -
Https-pages still in the SERP's
Hi all, my problem is the following: our CMS (self-developed) produces https-versions of our "normal" web pages, which means duplicate content. Our it-department put the <noindex,nofollow>on the https pages, that was like 6 weeks ago.</noindex,nofollow> I check the number of indexed pages once a week and still see a lot of these https pages in the Google index. I know that I may hit different data center and that these numbers aren't 100% valid, but still... sometimes the number of indexed https even moves up. Any ideas/suggestions? Wait for a longer time? Or take the time and go to Webmaster Tools to kick them out of the index? Another question: for a nice query, one https page ranks No. 1. If I kick the page out of the index, do you think that the http page replaces the No. 1 position? Or will the ranking be lost? (sends some nice traffic :-))... thanx in advance 😉
Technical SEO | | accessKellyOCG0 -
Does removing product listings help raise SERP's on other pages?
Does removing content ever make sense? We have out of stock products that are left on the site (in an out of stock section) specifically for SEO value, but I am not sure how to approach the problem from a bottom line conversion stand point. Do we leave out of stock products and hope that they turn into a conversion rate via cross selling, or do out of stock products lower the value of other pages by "stealing" link juice and pagerank from the rest of the site? (and effectively driving interest away) What is your perspective? Do you believe that any content that is related or semi-related to your main focus is beneficial, or does it only make sense to have strong content that has a higher rate of conversion and overall site engagement?
Technical SEO | | 13375auc30