Hi HN! I first posted CleverDeck about 2.5 years ago and received a lot of support and encouragement from the community. Since then, I've built it out into a more general purpose spaced repetition system and added seven new 3000-word frequency lists for major languages. I launched the big update today (http://cleverdeck.com/blog) and would love to once again submit it here for feedback. Cheers!
First, I was a long time power user of Anki and don't mean to disparage it, but I think CleverDeck has a much better user experience and design.
Second, we've spent a tremendous amount of time putting together high quality frequency lists (mostly for languages right now) that incorporate professional imagery, native-speaker audio, transliterations (where relevant), and example sentences. Usually, making or piecing together your own decks and cards is the most time-consuming part of using an SRS - and user content is often of dubious quality.
In terms of the spaced repetition, though, CleverDeck and Anki actually use the same algorithm.
> we've spent a tremendous amount of time putting together high quality frequency lists
I've recently started using Anki, but have only found it useful for taking a mental model that I sat down one weekend and learned, and re-enforcing it.
How valuable are these frequency lists for learning something entirely new? If I wanted to use them to say, memorise the api exposed by ActiveRecord, would they be useful, or would they ask me a series of questions that I couldn't yet really put in context?
What do you think is a fair price for one of our 3000 word language decks? In the previous iteration of the app, people showed themselves willing to pay $15.
Repeat revenue makes it easier for me to make a living as an indie dev, so I wanted to give this model a try while charging what I think is a fair price. I'll keep an open mind about it and adjust if enough people want it. Appreciate the feedback!
> Repeat revenue makes it easier for me to make a living as an indie dev
Of course it does! "And a pony."
You have to realize that most people are exceptionally leery of subscriptions as they are open-ended long-term commercial relationships with some random entity. There gotta be rock-solid clear benefits for a subscription to be justified. There are none in your case. What you have is a product, not a service. A subscription makes no sense whatsoever.
With regards to the fixed pricing - 3000 word pack is not the best option. What's more useful from the language learners perspective are packs that cover specific domain area - foods, kitchen utensils, car parts, body parts, weather elements, household tools, etc.
"3000 words" = "Unclear what you are getting"
"100 words, Garage" = "I may be able talk to the mechanic"
If you have smaller packs, you can sell them at few bucks a piece. You can try and experiment with discounted bundles of packs. And you can also try an all-you-can-eat option with a monthly access (yes, subscription), but as one of the options, not the only one.
PS. Have an option of switching off photos on cards.
I don't think domain specific packs are very useful for language learners. You need a rather large vocabulary just to follow a normal everyday conversation. Basic conversation and reading skill are the foundation upon which you can build. You need to consume and produce a large amount of material in your target language if you want to master it.
Buying the language a few hundred words at a time won't get you to fluency as quickly as learning words by frequency until you have covered the most frequent ~80%. Only then does it make sense to focus on specific domains that are most interesting to you.
Actually, it does make sense to have packs that are structured in a particular way.
For example, I am learning German. I would rather have a pack with the 25 most frequent nouns, 25 most frequent verbs, and maybe 25 most frequent adjectives. Then I have a foundation of vocabulary to learn the dativ and akkusativ. I don't need the adverbs yet.
Also, if I'm learning the modal verbs, I'd like a pack of those.
Domain packs are not for those who are trying to master a language, but those who are starting up and merely need to pick up some day-to-day vocabulary quickly.
So I googled "And a Pony" and it looks like one of the first uses of it is from a coding horror page in 2006 where the guy is asking all websites to scale down properly for a phone and then scale up properly for a desktop. He's then ridiculed because asking such a thing from websites would be obviously impossible. (Remember: 2006.) It is now of course expected everywhere.
I just think that's interesting in the context of your use here.
Hi. I actually had it in mind to create something like this myself -- a mobile-first, nicer-designed version of Anki. But it looks like you already started on this, 2.5 years ago. Always the way..
It definitely looks nicer than Anki. It's also better for creating simple cards, something which I haven't really been able to do on the Anki app. The workflow for adding a new card is a great experience. I do think that creating my own cards is an important part of the memorisation process. I've downloaded pre-made cards before, but the stuff I use in conversation is almost always the stuff where I've created the card. Those are the words/phrases that leap to mind.
One issue - my recorded audio didn't play when I reviewed the card (either on the front side or the back side). Maybe a bug?
Based on just 5 minutes of experimentation, I expect that this will become my "on the go" card creator, while I'll keep Anki around for the time being because, when I create cards on my laptop, I can import Forvo pronunciations etc.
100% this. I solely use the unofficial AnkiDroid port because it's leagues beyond the awful desktop app.
Somewhat related: a lot of its horribleness comes from the fact that the UI is written with QT, which is equally horrible for Windows and Mac, and basically okay for Linux as long as your DE is KDE.
> a lot of its horribleness comes from the fact that the UI is written with QT
I would agree with this. It also doesn't appear that much thought went into how the application should flow. It really feels like a smattering of UI components were picked out because they could accomplish the goal at hand without much thought as to what would optimize the user experience. For example, the way you browse through cards leaves much to be desired. Even stuff like the size of the buttons when you're reviewing cards. It DOES feel like the kit used to handle the GUI limits the application.
One thing that I think Anki got right (and I got wrong in my first iteration of a spaced-rep app) is doing away with the visual metaphor of a card. The reason you put information on a back of a real-life flash card is so you can review the question without seeing the answer. The answer is obfuscated, but readily available (you don't have to go look it up). There is NO reason to keep the metaphor alive in an app. It is actually less optimal to animate a card flipping over when you want to see the answer. It is better for the answer to simply appear underneath the question so you can easily reference each aspect.
What do you dislike as a user about Qt apps on Windows or Mac? Granted I _am_ on KDE (CentOS) but Qt apps that I've seen on Windows or Mac seem almost native. VLC, Opera, Anki
What is so bad about the Anki UI? I have been using it for three years on Ubuntu Unity and I'm pretty happy with it.
I especially like how keyboard-friendly it is.
Thanks a ton for the feedback - very useful. And thank you for the bug report - I'll check that out (my initial guess is that you don't have the CARD DISPLAY set to play audio on that side right now - which is itself a problem for being confusing).
Re: monetization. It's freemium right now - 100 cards to try out the app, then three bucks per month for access to all the decks we've made. Completely free to create your own content.
I'll look into Forvo integration - that sounds useful.
Ah, just found the deck settings and corrected this. Thanks. I've played around a bit more and have some further feedback.
Something that is really key for me: When I create cards, sometimes I want the audio to play on the front, sometimes on the back (and perhaps sometimes both). But there is just one audio value and I need to decide whether I want it to play on the front or the back across the entire deck.
For example, card 1:
FRONT: Text says "écureuil", audio reads out "écureuil"
BACK: Text says "écureuil", audio reads out "écureuil", and there's a picture of a squirrel
That's good for vocab comprehension, but what if I want the converse exercise?
Card 2:
FRONT: Picture of a squirrel
BACK: Text says "écureuil", audio reads out "écureuil", and there's a picture of a squirrel
At the moment there's no way for me to create both cards at once. I would need two separate decks because it's one template to a deck. And I would need to make the card twice.
I realise this is jumping ahead to 'power user' territory but it's definitely a feature I would use.
I also started on something like what you described (mobile-first, nicer-designed version of Anki) a few years ago, but catering specifically to Japanese. :) Not yet launched though.
Working on it! There's no really good way to do this on the device, so I need to build out a web app. But this is the number one most requested feature, so stay tuned :)
Do I understand correctly that paid subscription effectively removes 100 card limit? The description on the Settings page says something about "unlimited access to all content," but I can't seem to understand what this means exactly.
PS. I'm probably not the only one who stuffs his wallet deeper in the pocket upon seeing an "auto-renewable subscription." I would strongly encourage you to add non-subscription way for people to give you some money. I like what you made, it's useful and I'd like to support you, but no way in hell I'm entering a recurrent $ committment with a vendor of any app that I may or may not be using few weeks from now.
I don't necessarily feel like I have nailed the business model right now - I'll see how it goes. More broadly speaking, I think the jury is still out on subscriptions in productivity apps. Everyone is trying to get away with it now - and it is very attractive as an indie developer.
For what its worth, I feel similarly to you. I buy year long auto-renewable subscriptions, immediately cancel them, and then resubscribe if I'm still using them next year.
Great work, I've been searching for something like this for a while, and never really liked the UX for Anki. I love the UX, very clean and easy to use. Tutorial / intro was the perfect length, gave me a quick run down on how to use it without taking too much time.
I almost skipped over this because I didn't know what "spaced repetition system" was. My only suggestion is to market it around what benefit the user will get out of it, and not SRS / or the algorithm behind it.
Yeah, yours is the second comment along these lines and I think you're right.
My thinking was that I wanted to differentiate it from the very gamey branding of apps like Memrise and Duolingo. "Spaced repetition system" sounds more serious, more academic - and I do want people to get the impression that CleverDeck takes the spaced repetition implementation as seriously as Anki. I think I can probably still do that while being more clear - thanks for the feedback.
As a very heavy Anki user, I would love to know what you don't like about the UI. I use Anki mostly as Ankidroid on a Note 3, but I add all my material on the Desktop (CentOS).
Quick feedback: flicking away a card takes too long (vertical distance) of a flick. I have to carefully drag it from top to bottom for the card to flick away. Wish it was more sensitive. Perhaps 30% of the screen height is enough of a threshold. Otherwise, thank you. Going to be using this!
If I swipe long enough vertically, the card does seem to go away, so I figured it's a "come back to this later" feature. Didn't realize could do up/left/right to mark status, that's handy!
I love spaced repetition but serious question - what can you give me that Anki doesnt. Yes, I know Anki's interface is ugly but it is open source so I can actually change things and know how the spacing algorithm works. etc.
You can change most variables in CleverDeck's algorithm (the same one Anki uses), as well.
If open source and super customizability are important to you, Anki is probably a better choice for you. For many people, I think having the ready-made decks and a nice interface is more important. And CleverDeck does have some features, like three-sided cards and siblings, that I personally loved in Anki.
Personally I don't think super cuztomizability is that important for me. I prefer to have an out of the box solution with ready-made decks but being able to create my own decks is definitely a nice plus. My favorite aspects going through the app (learning Turkish) is the beautiful and intuitive design. Spaced repetition might sound a bit daunting, but completely got it once I started using the app. Well done, this is definitely something I will use daily and that will improve my learning.
Sometimes doing one thing well is all you need. Anki is like the assembly language of flashcards apps, higher-level solutions can be very handy too. (Speaking as a multi-language learner.)
The UI is great andI love that each one actually has example sentences, but the text isn't selectable and you can't tap to get the definition of other words in the example sentences.
The great thing for me is that there are great materials for lesser studied languages - I've used it for both Turkish and Persian, which ive really struggled to find.
Actually really quite well done, got it after the first posting, and quite enjoyed using it! Still do from time to time, the interface is seriously meditative :)
First, I was a long time power user of Anki and don't mean to disparage it, but I think CleverDeck has a much better user experience and design.
Second, we've spent a tremendous amount of time putting together high quality frequency lists (mostly for languages right now) that incorporate professional imagery, native-speaker audio, transliterations (where relevant), and example sentences. Usually, making or piecing together your own decks and cards is the most time-consuming part of using an SRS - and user content is often of dubious quality.
In terms of the spaced repetition, though, CleverDeck and Anki actually use the same algorithm.
One piece of feedback: The toolbar UX in the deck is confusing as it's inconsistent with every other iOS app. The 'back' button should be on the left and the 'options' button needs to be on the right.
Yeah, my mental model for the app has been that the main list of decks "lives" in the center, the settings live on the right, and the actual studying happens on the left (you'll see that it also slides in from the left - mirroring the exit animation).
You're not the first to mention this, though. I should probably change it.
How much effort is it to add a new language? My wife is trying to learn Serbian using spaced repetition and had an app to create her own flash cards with her own pictures and sounds, but it's very laborious
Have you heard about Memrise? It is site/mobile app dedicated to learn new language using SRS.
Apart from official content, most of the content is contributed by bilingual users. So I don't know the quality/accuracy of them, but the result is that they have lots of courses.
Yeah, I think you're right. The tagline can be improved.
I'm kind of doing a balancing act right now - I've built out CleverDeck to be a fully functional SRS that could be used for any subject, and I want to convey that. At the same time, people primarily use it for learning languages and I also want to communicate my dedication and efforts toward that particular niche.
I'm sure I can do a better job with messaging. Thanks for your feedback.
I mean, yeah, Quizlet is massive. I'll never compete with their user-content library and community. They've never had a particular focus on spaced repetition, so I don't really see them as a big competitor.
You can jump ahead in any deck by holding the red trash button. Since the decks are ordered by frequency, you could probably jump ahead 1000-1500 words and then just get rid of any words you already know.
Otherwise, you can always just create your own cards. If you do that on top of one of CleverDeck's decks, it will autocomplete if we've already made a card for a particular word.
Bandwidth shouldn't be an issue here, although this is a side project (so I realise money is limited), from experience you can send and receive roughly 200-400 images per second for every $5/month you spend. I don't know your precise usage count, but I would be surprised if you were to get that many images downloaded every second.
Right now, the import is pretty rudimentary. It just parses and creates a new deck.
I've been working on - and now will redouble my efforts based on the feedback here - a web app that will make creation and import much easier and flexible.
Not right now - I quite like using it as a universal app on my iPad. To be honest, I'm not sure how I could significantly improve it with all the extra space.
Thanks for confirming no plans. I've lost track of the number who have promised it for years and never delivered. This looks like a nice app for iPhones, but its not useful for serious iPad work.
In case you revisit this decision in the future, using the extra space isn't the issue. The "improvements" I'd be looking for on iPads would be much more mundane: support for the basic stuff like using the iPad keyboard, supporting split screen and landscape orientation, and text quality that doesn't look like it's from pre-retina days.
Incorporate a browser, perhaps? I am a student, so when I use Anki for classes I generally have to look up that idea to get the concept down. Would be pretty cool if you could have that integrated.
First, I was a long time power user of Anki and don't mean to disparage it, but I think CleverDeck has a much better user experience and design.
Second, we've spent a tremendous amount of time putting together high quality frequency lists (mostly for languages right now) that incorporate professional imagery, native-speaker audio, transliterations (where relevant), and example sentences. Usually, making or piecing together your own decks and cards is the most time-consuming part of using an SRS - and user content is often of dubious quality.
In terms of the spaced repetition, though, CleverDeck and Anki actually use the same algorithm.
Yes, in the fourteen days since you created your account, you might not have noticed that “Show HN” is quite common. I don’t know what the guidelines are, but perhaps they are outlined in the FAQ.
