About three years ago I built and launched Domain Pigeon, a web app that listed available web 2.0-style domain names each day [1]. The site did fairly well, but it was my first foray into web apps and I lacked the experience to grow it into something bigger. Eventually I moved on to other projects and closed Domain Pigeon down, which has been one of my biggest regrets.
My original vision for Domain Pigeon is what Lean Domain Search now is: you type in a search phrase and the app would pair it with hundreds (in this case 1,000) keywords to generate domain names and show you which were available. I lacked the technical skills back then to do bulk domain search quickly, which is why I settled on simply generating web 2.0-style domain names. Fast forward a few years and I've picked up those skills so I decided to take some time off my other apps and finally build this tool.
Hope you guys like it. Let me know how I can make it better.
How are you bulk checking now? I've been in the industry a long time and it's always been a pain in the ass to do anything in bulk. I've had lists so big registrars didn't want to run them for me :( Using APIs really isn't an option when you have a million names and 1 per second limit. Automating a bulk checker, most of them disable/play with it after a certain number of times (so even at let's say 500, i might be able to get 20 tries in there before they mess with it).
The best solution I've found is checking against the zone file, but I am curious what you're doing.
Based on a few tests, I think he's just doing DNS queries and assuming any NXDOMAIN response means a domain is unregistered, which is not always the case. Sure, users figure out the domain isn't available once they try to register it, but by then you've already dashed their hopes.
It's fine to use DNS queries as part of the checking process, but if you get a NXDOMAIN, you need to follow that up with some kind of check against a WHOIS server or root zone file (which is slower and harder to work with).
I'm guessing zone files. If you load zone files into something quick to search like an indexed DB table, memcache, etc they are super fast. They are also inaccurate and give you false positives since a domain can be out of the zone for a multitude of reasons.
You can do the initial search in the zone file and then if it doesn't appear there you hit the Verisign registry whois (for .com and .net). The amount of times the domain doesn't appear in the zone file is nominal.
Good answer, don't give away your secret sauce.
Having said that, your magic isn't working too well. I searched for http://www.leandomainsearch.com/search?q=apps
and literally EVERY SINGLE GREEN hit isn't really available after double-checking it. Seems like you'd need a 2nd secret sauce to further distill whatever the 1st run returned.
Hey -- there was a production bug that caused every result to come back as registered when you clicked "Double check availability". I fixed this a few minutes ago. Sorry about that.
I really like it. This proves there's still an immense number of viable domain names out there for startups.
Since you asked for some feedback, here's my two cents: I'd like to be able to sort the results by having my search term at the front or at the back of the domain name.
Highlighting the difference between verbs, nouns and adjectives is also something I'd like to see, taking a page from the book of http://impossibility.org/
Domain Pigeon was a brilliant site, I mourned its loss. It's wonderful to see that you've built a new site.
The execution is simple and easy to use. I'd be conservative when it comes to adding features or futzing too much with the UX. I find the color scheme very readable, although of course that might be different for some people. Thumbs up from me.
A post-search filtering mechanism might be useful, although of course you'll want to balance the benefit of a feature like that with the impact it would have on the elegance and simplicity of the design.
I've considered adding Domain Pigeon-style domain names to a separate section of Lean Domain Search, but as you note there's also a lot of value it keeping it simple ("do one thing well"). So we'll see what people want.
Agreed on the post-search filtering. Other than alphabetical and by length, is there any other way you'd like to filter it?
I'm a fan of the filters that NXDom provides: http://www.nxdom.com/
I suppose it's as much about sorting as it is about filtering, actually. I particularly like their ability to sort by length and readability.
Suggestions (I may be wrong):
1) I don't like very dark sites, yours look nice, but anyways...
2) I have no idea how you check the domains, but: In case you get your results one by one, you should display them while they were loaded (instead of that "loading" screen)
About three years ago I built and launched Domain Pigeon, a web app that listed available web 2.0-style domain names each day [1]. The site did fairly well, but it was my first foray into web apps and I lacked the experience to grow it into something bigger. Eventually I moved on to other projects and closed Domain Pigeon down, which has been one of my biggest regrets.
My original vision for Domain Pigeon is what Lean Domain Search now is: you type in a search phrase and the app would pair it with hundreds (in this case 1,000) keywords to generate domain names and show you which were available. I lacked the technical skills back then to do bulk domain search quickly, which is why I settled on simply generating web 2.0-style domain names. Fast forward a few years and I've picked up those skills so I decided to take some time off my other apps and finally build this tool.
Hope you guys like it. Let me know how I can make it better.
[1] http://www.hnsearch.com/search#request/all&q=domain+pige...