What Anki is missing especially in learning Chinese

For some languages with complex and visually similar characters — like Chinese — it’s very painful to recognize and efficiently review flashcards.

I’ve tried many Chinese apps, tools, and websites: Memrise, Learn Chinese Characters by Heisig, many apps, online courses, and of course Anki. There are some characters I’ve made so many mistakes on that they’re tattooed in my brain — but others I still get wrong, and it makes me angry. Out of frustration I break my streak and retreat to YouTube videos, reading books, or something else entirely. But I still make mistakes.

I’ve noticed these patterns:

• Sometimes I remember a hanzi in a particular word, but I don’t remember its meaning or even its pronunciation in a different word
• What I remember for sure is that this is the character I keep forgetting — and I’ve definitely seen it many times
• Sometimes it takes 5–10 seconds to recall the meaning or pronunciation of a hanzi
• Sometimes I know 2 characters inside a word but mix up their pronunciation order in other words. For example: 世界 shì jiè vs 介绍 jiè shào
• Sometimes I know both characters but stumble on which sound goes where 世界 shì jiè vs jiè shì
• Some hanzi are genuinely very similar: 己 已 记 / 万 方 仿 芳 访 坊
• Sometimes in an example sentence on my flashcard I see something I don’t quite understand, and I have to click around to find that word or character — which breaks the flow

All these relationships between errors are difficult to capture in Anki. Even if you add audio, word meaning, an example sentence, cloze deletion, and a picture — you still need some way to map the relationships between characters: to connect the ones you confuse with each other and work on them together.

I tried splitting things into different decks:

• Radicals
• Hanzi lists
• Words (2+ characters)
• Sentences

This became a nightmare to manage. Reviewing so many isolated chunks of mini-knowledge, my brain kept struggling to know: is this a radical? A standalone word? Part of a bigger word?

I tried my own decks, HSK decks, most-common-words lists, SpoonFedChinese. The problem isn’t the decks. The problem is that Chinese is a multi-level language — without an alphabet, everything can be a building block of something bigger, or itself consist of smaller elements that lead to confusion. Radicals build hanzi, hanzi build words, and the same character can be pronounced differently depending on the word. On top of that, there are tones. And writing — though I’m not focused on that in the AI era.

For me, the most important skill is recognizing characters first, then their meaning and pronunciation — especially quickly. Some characters I recognize instantly: I know the meaning, pronunciation, and can even produce example sentences on the spot.

So this is my pain with Anki — or maybe my pain with Chinese. Anki works fine for languages with alphabets. But to say I’ve truly mastered a Chinese character, I need to, within 1 second: recognize it, pronounce it, name 1–2 words that use it, and produce 1–2 example sentences. That’s the bar.

Here’s the kind of thing I’m dealing with during review:

Front — a card showing 练习 and 弹钢琴 together:

• 练习: I can recall 练 in ~3 seconds, 习 in ~1 second
• 弹钢琴: takes me 10 seconds to recall the meaning, and 20 seconds to fumble toward the pronunciation — something like tán wāngqín instead of tán gāngqín

Back — the answer side with audio, pinyin, and meaning:

I tried adding tags, flags, and grouping similar words into collections — but during review I keep finding new gaps in my memory or my strategy. And there’s another Anki quirk that frustrates me: if you fail a card enough times (I think 7 in a row?), the card gets suspended. Then it just disappears until you happen to notice it somewhere and manually unsuspend it.

I know Anki is not made for learning — it’s made for reviewing. But some of these characters used to be easy at the beginning. The more similar characters I come across, the harder it becomes to juggle all 4–5 dimensions of a Chinese word at once.

Other learners often say: the more characters you know, the easier it gets. For me it feels like the opposite — it keeps getting harder.

I tried reading, but I still haven’t fully mastered words that I can recognize in Anki fairly quickly yet can’t pronounce fast enough in context. I tried watching YouTube videos and vlogs — that felt more efficient. I can pause, I can read, I can hear, and I can use the visuals to guess meaning. I started using Language Reactor, and if I watch simple, lower-level content I follow it fine. But with dramas or more advanced videos, even simple words I recognize in subtitles I can’t catch with my ears. Listening turns out to be its own separate skill.

I even tried a dedicated listening course — just repeating and listening, no reading or writing, purely training my ear to connect sounds to meaning. It helped me realize something: knowing Chinese means developing many skills across many layers, and they don’t automatically reinforce each other.

Right now I know around 1,000 characters and have genuinely mastered maybe half of them. The rest I either confuse with similar ones or recognize too slowly — and the ones I don’t use regularly just quietly fade.

I started building a simple app focused exactly on this problem — distinguishing and fixing confusion between similar characters. It uses Anki as a data source and adds games, exercises, and drills to help imprint the differences and build a strong recognition skill for visually similar characters and words.

You can read about how I built it here: Building My Own Anki Browser

I’m not trying to sell anything. It’s going to be free.