The friction problem
Reading Japanese online without tools means stopping at every unknown word, copying it, looking it up and returning. The interruption is so large that most learners simply do not read, which is the real cost.
Tools that reduce that interruption change the calculation entirely: an article that would have taken an hour of stop-start work becomes twenty minutes of reading.
Volume is what produces reading ability, so anything that increases volume is doing the main job.
What the categories do
The tools available fall into a few types, and they solve different problems. Knowing which problem you have determines which is worth installing.
| Type | Solves | Cost |
|---|---|---|
| Pop-up dictionary | Lookup friction | Reduced retrieval effort |
| Furigana injector | Unknown readings | Kanji may not be learned |
| Reader mode | Clutter and layout | None significant |
| Sentence lookup | Grammar parsing | Can replace parsing practice |
| Text-to-speech | Reading aloud, pronunciation | None significant |
| Saved-word export | Building a review deck | None significant |
The bottom three have almost no downside. The top three trade effort for volume, which is usually the right trade and is worth being aware of.
Using a pop-up dictionary well
The failure mode is looking up every word reflexively, including ones you would have retrieved with a second's effort. That removes exactly the retrieval that strengthens memory.
A one-second rule handles it: try to recall first, and look up only if nothing comes.
- Pause for a second before looking up; try to recall.
- Look up words that block the sentence, not every unknown word.
- Do not look up a word twice in the same article; note it instead.
- Export the words that recurred rather than everything you hovered.
- Keep reading; the point is volume.
Furigana, which cuts both ways
Furigana tools make any page readable immediately, which is enormously useful and has an obvious cost: if the reading is always supplied, the kanji is never learned.
The workable compromise is to use it on material above your level and to turn it off for material at your level, where the kanji are ones you should be acquiring.
- On for material above your level, to make it accessible.
- Off for material at your level, where the kanji should stick.
- Off when re-reading something you have already worked through.
- Use it selectively rather than site-wide where the tool allows.
- Check periodically whether you can read without it.
Checking dependence
The risk with any of these is a comfortable illusion: reading fluently with support and being unable to read the same text without it.
A monthly check resolves this. Read something at your level with everything switched off and see how it goes.
- Once a month, read unaided for twenty minutes.
- Choose material you would normally read with tools.
- Note whether the difficulty is words or grammar.
- If it is dramatically harder, reduce the assistance slightly.
- If it is fine, the tools are doing their job.
Turning reading into review
The most valuable feature of a reading setup is not lookup at all; it is capture. Words met while reading are the best possible review material, because they come with a context and a proven relevance.
Exporting them into whatever review system you use closes the loop between reading and vocabulary, which is otherwise two separate activities.
Be selective. Exporting every hovered word produces a deck of things you looked up once and will never meet again; exporting the ones that recurred produces a deck of words your reading actually requires.
Capture what recurs, not what you looked up. The second is a much larger and much less useful list.

