[HN Gopher] Show HN: Anki-LLM - Bulk process and generate Anki f...
___________________________________________________________________
Show HN: Anki-LLM - Bulk process and generate Anki flashcards with
LLMs
Author : rane
Score : 34 points
Date : 2025-11-02 14:04 UTC (8 hours ago)
(HTM) web link (github.com)
(TXT) w3m dump (github.com)
| rahimnathwani wrote:
| I love this. Two things I noticed and liked:
|
| - the README is extremely detailed and clear: all the commands
| are explained with examples and the _why_ to use each one
|
| - you're using Anki Connect to edit decks in-place, instead of
| trying to edit or generate an apkg file. This simplifies things
| and avoids issues such as needing to create custom note types or
| avoiding creating two note types with the same field
|
| When my son and I have discussed a topic in response to a
| question, ideally I would evaluate whether there's something he
| should remember forever and, if so, I would create one or more
| Anki notes for that piece of knowledge. But right now it's too
| much effort, unless I'm at my desk. Even then, I need to copy and
| paste card fields from a chat interface into the Anki UI. That
| means I rarely do it.
| treetalker wrote:
| I'm always apprehensive about "efficiencies" like this because
| the process of generating the cards contributes substantially to
| the learning and memory formation.
|
| Can anyone help me understand the opposing view better?
| jbstack wrote:
| I completely agree with your first statement and I try to hand-
| generate my cards as much as possible.
|
| On the other hand, hand-generation is very time intensive.
| Having _some_ kind of Anki card for a topic you need to
| memorise is better than having nothing at all. If LLMs help you
| write cards that you wouldn 't otherwise get around to writing
| then it can be worth it.
|
| As an example, I've always found Anki really effective for my
| language learning. But the bottleneck was always finding the
| time to find good quality sentences from sources like grammar
| books and then creating the cards. Now I ask ChatGPT to
| generate me a whole bunch of example sentences for a particular
| topic or grammar point that I want to master, and I bulk import
| them into Anki in one go then use AwesomeTTS to create the
| audio. These cards feel less personal to me because I've lost
| the benefit of having put in the hard work of creating them
| myself from source materials. But that's more than made up for
| by the fact that I'm now progressing through the topics I need
| to learn at a much faster speed. I'd rather know 1000 words
| reasonably well than 200 words very well.
|
| After a few repetitions I don't think there's much difference
| anyway between a card you generated yourself and one you didn't
| - the SRS algorithm sorts it out for you in the end. The AI
| generated one might just need a few more reviews/fails/hards to
| get to the same level of memorisation.
|
| EDIT: I should add that I don't blindly trust ChatGPT's output.
| My wife is a native speaker for one of the languages, so I
| always have her check the cards. For my other language, I run
| the sentences past several other LLM models and I only keep
| those that all of them agree are correct and idiomatic.
| watwut wrote:
| > But the bottleneck was always finding the time to find good
| quality sentences from sources like grammar books and then
| creating the cards.
|
| I solve that bottleneck be seeking better books,
| documentaries and movies.
|
| Then I skip the flashcards step.
| utopiah wrote:
| 100%, in fact it's like when you write a "cheatsheet" only to
| realize that now that you did dedicate some time to
|
| - write down what is important
|
| - present it in a condensed manner
|
| - verify that it does indeed cover only the topic you need
|
| ... then ironically enough you probably do not need it anymore.
| arcanemachiner wrote:
| In that case, you may be able to upload the deck for others
| to benefit from.
| xdfgh1112 wrote:
| I have 15k learned. It's a question of timing. Can time spent
| making the card outweigh time saved learning it? I would say
| yes. It's easy to spend too long making a single card. A
| compromise is to make a small card at first and improve it
| whenever you fail it.
|
| Personally I need some context in a card to hook it up to other
| things. Such as the sentence where I first encountered it.
| Without that I will often fail the card over and over and waste
| time - it would have been quicker to put some effort upfront
| making a decent card.
| xdfgh1112 wrote:
| There is also a meta level investment in your deck that comes
| from curating it by hand, and that pays off in long term
| motivation AND improves recall.
|
| I'm sure some people can knuckle down and learn an LLM deck
| with random words, but they'd be a minority.
| atahanacar wrote:
| >improves recall
|
| Citation needed.
| atahanacar wrote:
| > the process of generating the cards contributes substantially
| to the learning and memory formation.
|
| How is creating a card anything different than reviewing the
| card once? Anki is a long term tool, writing something down
| once isn't. The time spent creating cards is better spent on
| doing more reviews.
| thomascountz wrote:
| I'm curious about the effect of "hand writing" a card for spaced
| repetition. It sure feels like it helps me learn more effectively
| when I write high-quality cards myself, but n=1 in this case.
| Even when I use an LLM to help, I have never find the cards to be
| useful by default--same goes for trying to use other's decks.
|
| That said, what I'd really love is a better card writing UI. If I
| could simply edit the table when in the browse view instead of
| opening the form view, that'd be a big step up!
___________________________________________________________________
(page generated 2025-11-02 23:01 UTC)