Hiding Furigana in Anki: The Built-In Filters, and the Catch

The short answer. Furigana — the small kana readings printed above kanji — can be hidden on the front of an Anki card and revealed on the back with no add-on at all. Anki ships three filters for it (furigana:, kana:, kanji:), and hiding is a two-line template edit: readings go into the field in Word[reading] form, the front template gets {{kanji:Field}}, the back gets {{furigana:Field}}. The catch is that every card's readings must already live in the field in that bracket form before any filter can do anything. For a deck you typed yourself, the syntax was part of the entry work; for plenty of shared decks the readings simply are not there in that form. If what you actually want is reading recall as the default shape rather than a template project, that design difference is where this guide ends up. Disclosure: I build Piccard, whose Japanese cards carry readings on the answer side by default — more on that at the end.

What hiding furigana is for

Furigana does the reading for you. While the ruby text sits above 会話, your eyes never have to produce かいわ — they only have to confirm it. That is fine when the card is testing meaning, and quietly useless when the card is supposed to test whether you can read the word at all. Hiding the reading on the front and revealing it on the back turns each card into a real recall question: kanji alone comes up, you produce the reading in your head (or don't), and the back confirms it. If you are studying Japanese for reading — news, novels, subtitles off — this is the difference between a deck that measures recognition-with-training-wheels and one that measures the thing you will actually be doing.

The built-in mechanism

Anki's ruby support starts in the field, not the template. You type the reading in square brackets after the characters:

`` 日本語[にほんご] ``

Three field filters then decide what renders, and which one you use is the entire hiding mechanism:

So the two-line version of "hide furigana on the front, show it on the back" is literally: front template {{kanji:Word}}, back template {{furigana:Word}}. No add-on, no script.

One syntax footnote worth knowing before you type a hundred cards: the brackets bind to the text immediately before them. The manual's own example is 世[よ]の 中[なか] — with a space — because 世[よ]の中[なか] places the ruby for 中 above the の. Misplaced ruby above the wrong character is the standard symptom of a missing space.

The catch: the readings have to already be in the field

Every filter above operates on bracket syntax that lives in the note's field, and no template edit can hide a reading that was never typed into that field in the first place. Which is why "how do I hide furigana" is really three different questions:

You built the deck yourself. Then the readings are in bracket form only if you entered them that way — and if you did, you already know the mechanism above, because entering the syntax is the harder half of the job. The template edit is the easy part.

The deck came from somewhere shared. AnkiWeb's shared decks vary wildly: some have readings baked into a display-only field, some have them in a separate Reading field, some have neither, and some render ruby only because the note type's template hard-codes it. Retrofitting the hide pattern onto a deck like that means touching the note type, the field mapping, or the notes themselves — for a 2,000-card deck, call it an afternoon spent maintaining instead of reviewing. The same maintenance tax that pushes people toward alternatives shows up here in miniature.

You use furigana-hover add-ons. They exist, they work, and they solve a different problem (peeking during review) while adding one more dependency to reinstall whenever Anki updates. The built-in filters are the stable layer; anything else is a choice to carry more setup.

The pattern in all three: hiding furigana is not a setting you flip. It is a property of how the deck's data was entered, which you either inherited correctly or get to redo.

Reading recall without the template project

Now my own product, with the bias that implies: Piccard's Japanese cards are built so the reading is part of the answer, not something the front renders. The word comes up, and the back carries the meaning, the reading, and the sentence the word came from. Reading recall is the default shape — there is no template to edit because there is no template layer exposed to you at all.

The capture side matches. Open a captioned Japanese YouTube video, tap the subtitle line that caught you, and the cards on offer come from that line and nowhere else. Saving one keeps four things — the word, its reading, the sentence, the point in the video it arrived from — and when the card falls due, the review plays that point back. While watching, double-click a word and the offline dictionary opens with the reading attached. Uploaded Japanese audio or video files follow the same pipeline, transcribed in Japanese itself. Scheduling is FSRS-6 with a fixed 0.9 target retention and no settings page. The honest boundary: this is not a converter for Anki decks you already own. If your deck exists and works, the built-in filters above are the right tool for it; this is about the cards you have not made yet.

The short version

Hiding furigana in Anki is built in: readings entered as Word[reading], {{kanji:Field}} on the front, {{furigana:Field}} on the back, {{kana:Field}} if you want the reading-first direction. The edit is two lines; the real cost is the data underneath it, because the filters can only hide readings that were entered in bracket form in the first place — self-built decks paid that cost at entry, and shared decks may never have paid it. If you would rather the reading-live-on-the-answer-side shape be the default instead of a project, that is exactly the choice Piccard makes.

The free tier is 150 energy (about 150 cards), then prepaid top-ups at $5 / $10 / $20; no subscription, and topped-up energy never expires. To start: Piccard on the Chrome Web Store.

Related: Best Anki Decks for Japanese · Kaishi 1.5k: What It Is and How to Use It · JLPT N5 Vocabulary

All guides: Piccard Blog.