As a non-native English writer, I always rely on grammar checking and proofreading software. One of the things I hate most (and the reason behind creating Typely) is a tool that stands in your way. That moment when you spend valuable time constantly adjusting something so you can do your job properly.
Just a few months ago I was in the process of launching a new blog for one of the websites we own and I began a little research to see the state of such services. The result was disappointing and I will tell you why in a minute.
Being a computer programmer, I tend to waste time trying to understand the “backends” of everything I use…pure curiosity. I discovered that 99% of the online services use the same “engine” with the rest of them being somewhat interesting but often too expensive for my basic needs or way too verbose.
Out of all the services I had on my list, most of them were constantly reporting false positives. For example writing a sentence that contains the word “Typely” or “Youtube” resulted in a call to action accompanied by a red line under those words. I hate red lines and anything that shouldn't be there for that matter so I can't go on unless everything is clear but what do you do when the flags are erroneous?
Since I can talk from experience now, writing such a tool is really hard. Trying to make computers understand what is the intention of the writer is near impossible and the results are often low in quality or simply annoying to the end user which, in most cases, leads to the loss of your audience. They lose interest because nobody comes on your website to train your engine (supposing you offer such a functionality in the first place).
Spotting a hole in this process I decided to try a different approach. My idea was to create an application that only targets things it is certain of. Typely is not an artificial intelligence nor a machine that goes out barking at every red flag. We created a huge list of uses like redundancy, jargon, illogic, clichés, sexism, misspelling, inconsistency, misuse of symbols, malapropisms, oxymorons, security gaffes, hedging, apologizing, pretension, and many more. Your text is submitted to our server and matched against over one thousand such checks. The results are impressive and we've only scratched the surface because we still have a lot more to add; the list keeps growing every day.
As a testing process, I always copy pieces articles from reputable publications such as The New York Times or BBC to see how they match against our service. Quite rarely I can see a report on these folks thanks to their knowledgeable editorial staff which is expected.
Typely reports only one questionable result
Following is a competitor's test result from a sample text (nytimes):
13 false positives being reported at only 180 words.
Typely remains silent and happy with some notices usually resulting from quoting someone (I'm still trying to find the impossible regex that excludes matches within quotes safely).
Typely stands on its own, doing its job beautifully and silently. I rely on my own tool to guide my writing. At just 1 month since the initial launch, it is part of my writing stack and I couldn't be happier with the results. There is still a lot to cover but the road is clear for us: a sensitive, unobtrusive and reliable tool for any writer, newspaper editor, teacher, blogger or student.
Subscribe to Typely Blog
Get the latest posts delivered right to your inbox