Spellr.us needs a new dictionary

Spellr.us needs a new dictionary

Summary: One of the only Australian start-ups to present at the recent round of conferences in the US was Sydney-based spellr.us, which has launched a Web-based tool to check and monitor websites for spelling mistakes.


One of the only Australian start-ups to present at the recent round of conferences in the US was Sydney-based spellr.us, which has launched a Web-based tool to check and monitor websites for spelling mistakes.

(Credit: spellr.us)

Founder Kevin Garber, who is also founder and general manager of spellr.us parent company Melon Media, told bootstrappr that the self-funded start-up grew from the need of some Melon's customers wanting to keep their websites free from typos.

To use the system, website administrators create a log-in, input the URL of their site and set a number of options, such as how many pages they want to be scanned, how often the scan should run (from no scheduling right up to once per month), and what dictionaries they want to use.

So far only English is supported, with options for US, UK, Canada, and international variants, although users can also set the site to take account of Australian slang, technical jargon, medical terms and so on. The European set of languages (Spanish, German, etc) are "under construction".

Garber has quite a bit of experience in the online arena, having previously worked for local start-ups during the dotcom years, as well as the Australian division of internet giant AOL. He has also worked for a South African talkback radio station.

The immediate next steps for spellr.us will be to transition from beta into a live version and gain traction with paying customers. The site reveals that pricing is "still being finalised", and states that there will be a variety of different options, including a free plan for small sites and a corporate option for customised solutions. Spellr.us is also planning to offer a reseller program.

Garber said the firm's goals were to grow fast and extend its product range; the founders would not consider an exit until the medium to long-term time frame.

There is no doubt that spellr.us has potential. The start-up has come up with quite a unique idea that translates well into a Web service that bootstrappr can see quite a lot of people being curious about, especially large sites that publish a lot of their own unique content.

 Kevin Garber spellr.us

spellr.us founder
Kevin Garber

(Credit: spellr.us)

However, there are some massive roadblocks in the company's way.

Firstly, when we tested spellr.us with ZDNet.com.au, we didn't find the service that impressive as it's currently implemented. We set the "International", "Australian slang" and "Tech jargon" options and hit go. Spellr.us took quite a few minutes to come back with a result. Unfortunately the results weren't spectacular.

For starters, the estimated "50 seconds" timer on the query was inaccurate ... it took a few minutes. The spellr.us page refreshed several times during that period for no apparent reason; unusual in this day of sites built on AJAX technology.

When the service did eventually finish scanning, it listed a number of common technology words (virtualisation, VoIP, ITIL, PDAs, datacentre, whitepapers, XP), local company names (Telstra, Optus) and even ZDNet.com.au sister publications (CNET, BNET) as among the 3,000 errors found. In fact, none of the top 20 errors were, in fact, errors.

(Credit: Renai LeMay/ZDNet.com.au)

Drilling down, it did find a few legitimate errors, such as "weathe", which should be weather. But the vast majority of the problems found by spellr.us were correctly spelled words. We had a similar experience when we tested spellr.us on wotnews.com.au.

Sure, ZDNet.com.au is a niche site with a lot of technical jargon, and spellr.us does allow you to add words to its dictionary. But very few sites are written in plain English; most will have company or industry specific jargon. In its current form spellr.us didn't "just work" for us, which is our crucial test for any start-up.

The second problem is the site's business model. The fact remains that most large companies that operate websites (even banks, for example) employ in-house copywriters or editors to put up content and check for errors. Sometimes this sort of content is published by the website development team, or a marketing/communications specialist.

Sure, having the ability to automatically check for spelling mistakes is useful. But by and large, when most large companies already have staff members paid to do the same thing, it's hard to see them paying extra for such a service. In short, we haven't seen much of a problem with typos on websites.

Spellr.us' service could be ideally rolled into a Web hosting platform as an optional add-on for hosting companies to sell to large or small businesses or individuals, but even then we can't imagine people paying a lot of money for such a service in the short term. It would likely be a free add-on to a content management system; Australian Web hosting businesses typically offer dozens of such free add-ons (although we haven't seen anything quite like spellr.us just yet).

Sorry spellr.us. We know Jason Calcanis gets, loves and wants your offering. But bootstrappr thinks you've got some work to do.

bootstrappr opinion: BUST

A fan of the service in the TechCrunch50 demo pit:

Topics: Start-Ups, Enterprise 2.0

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • You're right, but also wrong

    You're right, it picks up a lot of "correct" misspelled words on first scan. I imagine this will improve quickly (and easily with a bit of time from spellr.us) but the ability to quickly create your own custom dictionaries negates this early weakness with a just a few minutes work on your own part.

    For the value it adds to your content that few minutes is well spent. Once you fix a common word on your site that spellr.us detects as an error its fixed permanently, you won't keep getting told about it in the email reports.

    Wanting it to "just work" is all well and good, but if you consider in the context of the effort required to build a high quality content-driven website nothing "just works" with no initial effort. I imagine the CMS behind ZDNet didn't "just work" when it was first installed without some time and effort spent on it.
  • Had a similar experience - nothing training can't fix

    I found a lot of errors as well after a first pass - a lot of noise. Updating the dictionary does the trick, it's akin to training voice recognition software IMO.

    Once this is done I found it provides a very useful service to pick up embarrassing tpyos - oops I mean typos. ;)

    Also it doesn't detect grammatical errors (although this is mentioned as a possible upcoming feature - http://spellr.us/faqs.php ).
  • iterations

    Hi All

    Thanks for the article Renai and for the other feedback from the comments.

    I actually agree with your comment that it should "just work" and we are aiming towards that.

    Also we have implemented AJAX refreshes in the new version. The new version is a few weeks away and I think you will be pleasantly surprised. One of the changes is a very smart algorithm that works out what errors are most likely to be genuine typos even if you don't have a custom dictionary in place.

    Regarding the business model:

    We actually signed up one of Australia's largest telcos for a 3 month paid pilot a few months ago - they have a 3 week staging process for their content and typos still get through.

    If you look at our site http://2xmad.spellr.us/ you will see how many sites do have typos.

    So we believe that people will pay - but the final proof will be when we move from free to paying.

    Thanks again - great to have someone focussing on Aussie start-ups.

    Kevin - founder spellr.us
  • CMS, just working etc

    You're definitely right -- building complex websites is hard. In my articles I have taken a choice to use a level playing field evaluation of startups (we have a standardised questioning process, for example), but I can certainly see the potential of a few services that aren't quite perfect yet, Spellr.us being one of them.

    Most websites are a work in progress.


  • Dictionary

    Interesting to see a couple of comments along these lines. I'll definitely go back and keep playing with the service, adding words in and testing it etc as it evolves.


  • Start-ups etc

    No worries Kevin!

    It's a great credit to you that as the founder of a startup you will engage on forums such as this; it gives me a lot of confidence in how you guys will go in future.

    Looking forward to seeing what you guys do next! And as always, I am very happy to be proved wrong about the boom or bust call. Actually my main hope for Australian startups in general is that they succeed, put Australian technology on the map and build the next Microsoft or Google -- right here Down Under.

    There are certainly enough talented people in the local market for that to happen, and the local start-up scene is certainly hopping at the moment.


  • I think you were a little harsh

    I spent many hours last week cleaning up three websites with about 800 pages between them that we maintain. I consider it time well spent.

    Because we had a lot of jargon and people's names listed I ended up with a dictionary of more than 1200 custom words. I think perhaps 100 of them should hav been in the standard dictionary. I think the service is well worth paying for in our case. I blogged about it here:


    I am not affiliated with the company except that I plan to become a paying user.

    Sean Murphy www.skmurphy.com
  • new version


    Would love you to give an updated review - the system has moved in leaps and bounds since your last review - and we are keeping development up at a rapid pace.