Should you pay for Dragon Anywhere?



I love dictation. I love how quickly it enables me to write and, in turn, the time I save to do things other than working – like being with my family or bingeing on the latest series of “House of Cards”.

As wonderful as it is, though, I find it incredible that we are still tied to a desktop application to get the best results. Obviously, if you are dictating into your PC or Mac then Dragon has the ability to learn from its mistakes and build an increasingly accurate profile of your voice and writing style. The processing power required to do this (and to physically store the large file containing all your dictation data) is beyond the capacity of most mobile devices.

Until now.

A few weeks ago, I took delivery of a shiny new 12.9 inch iPad Pro. Once you get over the initial shock of a tablet that is literally bigger than your head, you are struck by the sheer power of this thing. The A9X processor Apple has put in this device (combined with 4 GB of RAM) is desktop class and, boy, can you feel it in use. Along with its recently released smaller version, this is a tablet that I have no doubt could comfortably run an iOS-optimised version of Dragon.

So why won’t Nuance make one?

Instead, they’ve given us Dragon Anywhere. On the surface, it’s great – super accurate, continuous dictation finally available on a mobile device. But just like the long neglected (but still perfectly usable) iOS Dragon Dictation app, all the recognition processing is done remotely. It’s the same as using Siri or Google Now – you speak, a server somewhere on the planet thinks about what you said and, a second or two later, the results are whizzed back to you. The big advantage of Dragon Anywhere is that unlike all of those examples, there’s no 30 or 60 second limit on your dictation – you can just keep talking, as you would in the desktop version.


One big upside is you can run this on pretty much any mobile device that supports it – both iOS and Android. The downside? All of that power I talked about in a device like the new iPad Pro is going to waste. The sad reality is that desktop PC sales have been declining quarter on quarter for many, many years. While Microsoft certainly doesn’t need to worry about the state of Windows just yet, there’s no question that people are moving to Macs or even mobile devices to get things done.

I love the idea of one device that can just about do everything. I’m not a gamer or too heavily into video editing, so my mobile needs are simple – I want something lightweight with a great screen that is capable without needing liquid cooling. I have a Mac desktop (also running Windows under Parallels) for the heavy lifting back at my desk. I’ve owned several Microsoft Surface devices (including the latest Surface Pro 4) and every one of them has been riddled with hardware problems, software issues or both. As ideal a portable Dragon setup it seemed, I finally gave up in frustration on that avenue recently.

The iPad Pro would be an ideal Dragon device for me on the road – it’s easily powerful enough, has plenty of storage and the screen is gorgeous. But Dragon Anywhere doesn’t take advantage of it and I think the reason for that is simple – money.

Dragon Anywhere is expensive. Really expensive. It uses the subscription model (something I have a sneaky feeling Nuance may shift to in the future for the desktop product, as well) and costs an eye-watering $15 per month or $150 per year. For a mobile app that doesn’t even interact with the desktop product (despite what Nuance claim, the synchronisation is minimal), the pricing is absurd. And this ties in to the reason why Nuance will probably never make a desktop-rivalling mobile version of the program that doesn’t rely on an Internet connection – they can’t sell it for enough.

Mobile apps are cheap. Could you see Nuance developing and selling a version of Dragon for iOS and Android that cost, say, $10? Not a chance. Not even $20 or $30 would cut it, and that would be considered expensive in the average app store.

So $15 per month it is, then. Is it worth it? It depends on how desperately you want to use Dragon on the move, but I struggle to see the value. Dragon Anywhere needs a constant Internet connection to work and, disappointingly, is little more than a dictation notepad. You can’t simply talk directly into, say, your mail app or the mobile versions of Word or Pages. Everything you say is translated within Dragon Anywhere and must be copied and pasted from there – and, of course, if the app crashes, it will take your dictation with it.

It’s also accurate, but only to a point. There’s no profile being saved, so the only thing it remembers (and syncs, if you have DNS Professional or Dragon for Mac 5) is your custom vocabulary. There’s no training of your voice or writing style going on. None.

So, it seems, the best way to dictate on the move is still to use transcription. That means having the full version of Dragon installed on a Mac or PC somewhere, but until Nuance rethink their mobile strategy, it’s probably still your best way of writing while you’re away from your desk.


10 thoughts on “Should you pay for Dragon Anywhere?”

  1. I’m testing it right now on a Chromebook that has access to the Google Play store and thus Android Apps. My problem is similar to what you have said. I would much prefer a local version or a cloud based one with a custom profile. I don’t care if I always need internet access, I just need a custom profile and global dictation, which should be easy for them to roll the dragon anywhere cloud service into a keyboard with SR.


    1. The mere thought of having to speak into a dictation box every time I was to try to use Dragon Anywhere in my apps is a dealbreaker for me. I can do basic dictation on my Android devices using Swype+Dragon, a 99 cent app. I can dictate into any app: Word, Facebook, Gmail… And even though the more advanced correction features aren’t there, I find it’s pretty accurate. I don’t get why Nuance would bother making an app that’s so expensive and yet so useless…


      1. I might have to try that one. Definitely not paying $15 a month after just purchasing the whole package.


      2. Well the reason why people want Dragon Anywhere is because of the ability to edit with your voice something that none of those speech recognition apps have. I write professionally and the ability to edit with your voice and navigate with your voice within a document is something that you just can’t give up once you have used it. I have used every single speech recognition app that is in the android store and every single one that is available on the Internet via the browser. None of them compare to any of the Dragon products. I do agree that Dragon Anywhere needs a bit of improvement but I most want it to be a keyboard with SR along with the editor that they have. I want to be able to use it anywhere on my mobile device and I want to be able to edit and navigate with my voice within a document.


  2. The 2017 Apple iPad Pro has more than enough local processing power for a full desktop version of Dragon but Google and Microsoft are now pushing AI NLP solutions. Dragon will be pushed out of the mobile market because it is not willing to do the R&D investment required.

    Sadly, my Surface Pro 3 and Pro 4 were too glitchy and unreliable for Dragon, not helped by severe processor throttling. I won’t risk the latest Surface because it is unlikely to be stable enough.

    So, although I have the latest ipad Pro which is stable and reliable, I have no voice recognition solution on it

    So I use a digital voice recorder in conjunction with desktop Dragon Professional Individual.

    Sadly, the latest version of ipad Pro although an ideal hardware platform for a Dragon desktop solution will never receive it. Siri has a Dragon engine but can’t do full scale voice dictation and recognition. So, unless Apple delivers an AI solution as a local App, that will exclude the iPad

    It is a real shame that a stable Surface solution is beyond reach. That leaves a HP or Lenovo competitor to the Surface as potentially viable especially when Coffee Lake arrives.


    1. It’s a real shame about the current situation regarding voice recognition on the iPad Pro – as you say, the hardware is more than capable. I’m not sure Nuance doesn’t want to invest in the mobile market – my guess is they are betting on the desktop solution having a long-term future. Either way, they would have to adopt a subscription model and the current pricing for Dragon Anywhere, even if it was brought into line feature-wise with the Mac or PC programs, is too high for most people to bear in the app market.

      I feel your pain regarding the Surface line. I went through a number of Pro 4 machines with severe reliability problems, much of which was software-based. They have sorted out it now though and I’m using one successfully with DPI 15 – the length of time it took Microsoft to resolve those issues, however, was unacceptable. Now they have started pricing the Surface line into oblivion, maybe third parties can step up, like you say, to fill the void.


      1. I should mention thay I purchased your wonderful book on Training the Dragon which transformed the usability of Dragon 13 for me. I now use Dragon Professional Individual 15 on a very powerful HP workstation and get excellent results, but it is not a portable solution.

        I would love a mobile tablet based solution but I have had terrible luck with the Surfaces and never succeeded in getting the numerous problems sorted out. I lost confidence in the products. They failed at the most critical times.

        My iPad Pros function robustly and reliably. They simply can be relied on to work. That means a lot. Now they are processor powerful and far more than enlarged mobile phones. IOS 11 will be transformative.

        Even a non command version of Dragon would be wonderful. Of course, there would be a lot of development work porting over to an ARM based processor.

        A friend is working on a deep machine learning approach system using 16 very powerful GPUs which learns from a single voice. It is far too time intensive to be tolerable because it is working slowly to learn his speech patterns, vocabulary and language structure. Already he is achieving exceptionally good results with familiar learned words despite his odd and indistinct accent. It involves a lot of correction and confirmation, based on feedback, but the learning development is astounding. At this point, using a USB microphone the system is delivering a lot of resilience and fault tolerance based on the patterns it has learnt. Variations in tone, speed, pronunciation and background noise are not affecting accuracy but the impediment is that it is uniquely tailored to one user.

        His theory is that the system in time will learn much more quickly from other users but it will require more reading input than even the earlier versions of Dragon. 10,000 words spoken 3 times is likely the minimum sample base. This would be unacceptable to most users but the advantage is that accuracy is almost infallible for known words although painfully slow. He estimates 32 GPUs, more efficient parallelism and wider buses would give practically instantaneous results.

        So, it is not a commercial approach but it may lead to commercial application if refined. The investment in training the system is huge and would be offputting to any prospective end user but is a true learning system. It is tedious, slow and hardware intensive delivering astounding accuracy for known words slowly. It is no good with names, proper nouns or technical expressions unless added in.


  3. I think Nuance simply have it wrong… an app doesnt have to be a continuous server side fed wifi monster to be worthy of a subscription payment plan.
    Plenty of apps are on a subscription plan of 3-5$ per month for an app that reside 100% on your device, they could easily do the same and save the server cost for themselves AND make continuous amount of money every month. By not having to pay for servers to deliver the work they could drop their monthly pricing to something reasonable and by making it a “keyboard” app they could have it work with most apps.
    Compatibility wise? They could just release it for the machines that actually can do the work like the iPad Pro and users would probably be happy to upgrade their hardware to get the best app wise.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: