hi, I took a subscription today for the export to Anki. I thought it would be an auto-sync to Anki, but it’s quite manual work.
Would it be possible to do have that auto-synchronization to Anki?
Anki connects allows you to have API calls if the user adds this addon:
This would be amazing!
Hi!
We just pushed out a big update to Anki Export (with screenshots and pretty good TTS audio).
Agree about the export/import being clumsy. I have prototyped with Anki Connect and it’s pretty easy to add that function, now that the Anki Connect interface has been documented on the Github repo. It will need a Webstore update, but I think this should be included in the next release.
D
Hi David, this is great news.
May I ask what TTS software you are using for the Thai?
It sounds much better than OS (Thai Kanya Siri) and Google TTS
If you could let us know that would be awesome!
Keep up the awesome work… I just became a pro-member
Regards
It’s a Microsoft API. We switched all languages over from Google because the voices are nicer. They are wiping the floor. We are actually training our own TTS model for English now so we can make more use of TTS for future features.
Hi David,
Thankyou very much for the response and the information. I do have a final query. In regards to Anki(+CSV) download option.
Is there are reason you have it all coded into one field? (text)… by that I mean that it becomes very limited in that we can’t get at all the different fields such as to control the audio, change card layouts.
I’m struggling to see the logic behind putting it into one field (images, close text, audio)… such things as auto audio play ® thus don’t work in Anki… you can’t add the audio field twice to auto-replay which is often a nice thing to do in Anki when listenining to audio.
The CSV download option on its own is great as you separate them out into different fields… it’s just missing the images/audio… but the Anki+CSV option although has the images and audio which is great it’s all somewhat bounded together…
i.e you might want the images to show on the back of the card after trying to guess from the audio what they are saying.
Each person has their own way of doing it.
I can get it to work by using just the CSV option and then combining it with the media from the Anki+CSV option but then manually matching up the audio to a new field is cumbersome and not easy to get in the right order.
Happy to work with you on this and test… essentially looking for the CSV option (where they are all in different fields) and the media reference added to the end in 2 different field (images, audio) and then the media files in another folder.
That would be perfect so we can import and match perfectly.
Don’t get me wrong I’m loving the work you are doing I just see this as a better long term approach (or to have both)
Let me know if I can help at all.
(I work online in the php/laravel/react/ecommerce sphere so I’m always tinkering about with stuff!)
Regards,
Hey. Thanks for this detailed feedback. Yes, “Anki(+CSV)” option wasn’t upgraded in the last update. So, here’s the todo list:
– For ‘basic’ Anki export, use special sound format ("[sound:FILE1.mp3]") instead of html5 audio tag. (We use html5 audio tag because we were experimenting with using dataURIs as audio src URL, to avoid having to copy files to the Anki media folder. It worked well, but AnkiWeb would only accept about ~2000 such cards before the storage limit is reached. With seperate media, no such problem.
– Add names of images/audio files to CSV output.
Just these two fixes? Should be able to do it tomorrow. 
David,
That would be truly awesome if you could do that.
“Add names of images/audio files to CSV output.”
This would be perfect! plus you also add the media folder as an option for the download (like you have with the other one)
That way you have everything you need to create literally any Anki card or format that you’d like.
Support wise it would be a smart move, as you can just direct anyone to that download so they can create whatever they like with Anki!
Then for your Anki(+CSV) one where it kinds of gives people a templated version (so they don’t have to do any manual stuff themselves) using the [sound:FILE1.mp3] is definitely the way to go.
I’d really appreciate you doing those updates!
(I just connected an API key through the TTS in Anki itself to Microsoft Azure… didn’t know you could do that!)
Again thanks for being so responsive, I’ve signed up as a pro member but if you also have a donate button… happy to contribute.
This is such a marvellous resource.
Looking forward to that update!
Regards,
Hi Sven,
Sorry, taking a bit longer than expected. Og added some code for downloading exports in multiple parts, he went to see his mother today, I’ll wait till he gets back tomorrow before editing that code.
D
Hi David,
No worries at all, thanks for keeping me in the loop!
Hey. Added the extra cols in Anki+ export, Anki+ can export with media now, changed ‘regular’ Anki export to export [sound:FILE1.mp3] format. btw, ‘regular’ Anki export has a nice set of html classes that can used to customise appearance. Should document these in instructions.
Marvellous!! Exactly what was requested… Thank you very much David.
There was just one small thing (lol)
The first column of the CSV export is “Item Type” which is only either Phrase/Word
Anki uses that first column/field as a Duplicate identifier through its app/warnings and for importing. (so basically all rows from the CSV upon import are seen as duplicates as it’s looking at that first field)
The first column could really do with being a unique (ish) type key… i.e the “subtitle” column
There are manual ways around it, i.e opening in a spreadsheet, manipulating, exporting each time also ignoring that field on import (but this has small niggly knock on effects on the imports settings each time) so they are rather cumbersome.
It makes much more sense to have the first column as the “Subtitle” field… could you change the order?
After that we are all set… what an awesome resource… really
Hey, yes, this was important to do. We just made the change. The first field is now a key which we use internally, based on the lemma and language of the saved word (for phrases a different kind of key is used)… it means if you change the colour of a word in LLN, a new item should be exported to Anki, but this will be detected as a duplicate card… I suppose you have the choice to update the data of the existing Anki item.
