Subscribe
Sign in
TypingDNA Verify (2FA) — Replace SMS 2FA codes with better UX, just type 4 words
Want to replace annoying SMS 2FA codes? We verify your users by the way they type 4 words. This improves UX significantly and reduces costs by an order of magnitude. Btw: you can get the solution free of charge for a limited time. Check the website for more.
51
Replies
@chrismessina in our tests our solution is much much faster, not sure why but a 6 digit code is harder to remember in one step so you'll need to go back and forth at least 1 time (typical use), you also need to find the phone, make sure it's unlocked, look for the right notification, wait for it to arrive, make sure phone doesn't go to sleep while typing the digits/alphanumeric codes in the app. For example, above you typed a sentence of 10 words, I bet it took you less than 3 seconds. Now think again about how long it takes to find the code in your phone and type it. Wdyt, is it more than say.. 2 seconds?
Top Hunter
Yeah, I submit that I'm not normal. I have an Alfred Workflow that will grab the Sox digit code from iMessage and put it on the clipboard, so it's somewhat irrelevant what the content of the code is... But deviating from the six digits would likely break my current solution! 😅
@chrismessina I think the biggest issue within product community in general is that none of us is normal :) yet our end-users are.
@raulpopa it's weird how I was just looking you guys up for your product's use for open banking SCA-everyone talks about the pain/friction of SMS OTP when making direct account to account payment in open banking but they talk like solutions like these don't exist.
Curious whether you're cooking up anything to boost adoption among the open banking adopters?
PS: When I saw you product hunt after coming from your website, I thought you were retargeting or stalking, hehe...such a coincidence.
TypingDNA and ThreatMark were on my list.
This looks like a no-brainer for virtually any company that's offering 2FA.
Excited to see the product live so I won't have to go find my phone and deal with those annoying texts or Authy while browsing from my desktop.
Congrats folks!
@andrei_negrau Thank you, Andrei. Ping us when the time comes for Cartloop to implement 2FA.
@darren_chait Thank you! Greatly appreciated! This goes for you as well, congrats for the great work your're doing at Hugo!
@george_brata Merci! I would be curious to hear how helpful you think typing four words are instead of SMS codes (from your mobile phone to your laptop).
@wisepythagoras Verify typing verification doesn't work on mobiles yet (at all), we just send the SMS code on mobiles at this time. However TypingDNA has developed other technologies that handle mobile typing, but swipe typing is a thing we didn't yet solve publicly. We have a work in progress there. It's a quite hard problem to solve, probably will not have that accuracy of the standard typing. One thing a mobile integration of our Authentication API can do today is to ask people to type in a standard way or display a custom keyboard that has swiping and predictive typing blocked. Does this answers your question?
How does it work for cross-device detection?
I might have a different pattern on my mac and pc. Also a very different pattern in mobile and ipad.
@freakyarch thanks for the question: the solution is device agnostic within all desktops and separately within all mobiles (but mobile solution is not yet available). btw, your typing pattern will not be significantly different on a mac vs a pc.
@benoit_chambon Sure! Keep it in mind for whenever you might need 2FA for Koinju.
Having used TypingDNA in the past, I can say this is amazing! Keystroke Dynamics does need to have more use-cases and it's nice to see it heading in that direction.
Good luck :)
Wow - this is extremely creative. You could be onto something huge! The future of verification. Exciting stuff, I'm looking forward to keeping up with your progress.
@glen_creaser We are doing our best to do so! Do you think typing 4 words is quicker than having to look for your phone and type a verification code on your laptop?
The best product from Typing DNA so far. Congrats! On the road to becoming a unicorn :)
Godspeed!
How do you accurately capture keystroke dynamics and ensure consistency? May be quite awful if this malfunctions and user finds it frustrating to just get it passed. Great potential, really like this
@daryllman Great question! The technology is trained to recognize people when they type on different devices, different keyboards, etc, but it's not perfect. Sometimes it fails to recognize same user, especially if they type in a very odd position, with just one hand, or intentionally trying to disrupt their natural typing behavior. Luckily this happens rarely enough and when it does, we just send the good ol' OTP code to their phone. The main technology behind Verify is already in production with a number of clients and we're yet to see one client leaving. We have 0 churn, which I think translates in good overall satisfaction, and less frustration on the user end.
@raulpopa This sounds really awesome! Would really remove the hassle of 2FAs (literally a pain in the ass). All the best! Looking forward.
What level of uniqueness can you discern from someones typing and how can you accurately match that across multiple devices? One of the most popular uses of 2FA is when signing in on a new device and I imagine ones typing "fingerprint" would vary quite a bit.
@devin_alldrit1 this is a very deep question, I can either go with a long answer or a short one, there is no middle way. So, here's a short(er) one: user's typing dna measured on 4 words of ~28 chars with 3 initial enrolls (our standard) was able to produce 0 false positives in our tests (with thousands of samples). What we do in addition for Verify, since we produce the word combinations ourselves, we have learned a new algorithm for scoring text complexity/strength, that way the words you get to type add to the initial uniqueness of the system making it highly reliable and very hard to fool. It also becomes much stronger with time. That said, different devices work almost as well. We've tested it extensively and proactively recorded same users on multiple devices on purpose to support this use case in particular.
@slava_bobrov like with any biometric based solution there is a chance to get a false negative (also called false reject), as well as a false positive (false accept), but these are very rare, we're using a state of the art typing biometrics authentication engine developed over the last 4 years with really small potential false matches. The solution itself uses SMS 2FA codes as RoT (Root of Trust), therefore it already starts with some weaknesses. To summarize: like all other biometrics, and like the standard SMS OTP solution, this is also not perfect, that's why we only recommed it to be used as a second factor.
TypingDNA