Understand what a word means here

You look a word up, read the definition, and still cannot tell which sense the sentence in front of you intends. That is the most common failure of a plain lookup: a dictionary definition is general by design, while your sentence is specific, and the gap between them is precisely the thing you are stuck on.

Capecho closes that gap. Ask what a word means in this sentence and it explains the exact sense the word carries in the context you captured — and, when the phrasing itself is the obstacle, what the whole sentence means. This is AI word meaning in context: not the word in general, but the word right here.

A word like 'figure' can be a number, a shape, a person of note, or a verb meaning to conclude. The general explanation lists those senses side by side; it cannot tell you which one is live in your sentence, because by design it never looks at your sentence. Choosing the right one is a separate question.

The in-context explanation is the part that answers it. It reads your specific word-and-sentence pair and resolves which sense applies, so an AI contextual vocabulary explanation does the disambiguation your eyes were trying to do — and explains the reasoning rather than leaving you to guess among the options.

Capecho keeps two explanation layers distinct, and the distinction matters. The word explanation — core meaning, part of speech, distinct senses, pronunciation, and a system-Dictionary handoff — is free and unmetered, generated once per word and shared from a public cache built from the word alone. For many words, seeing those senses laid out is all you need to know which one your sentence means.

The word-meaning-in-context AI layer is metered: 10 in-context explanations a day, free, with unlimited on Pro. It costs per use because every sentence is different and the result cannot be shared or cached — and hitting the cap never blocks the word explanation, saving, your Word Book, or review. Everything else keeps working, and the in-context count resets at your account's local midnight. (Pro also lifts the free tier's 200-saved-word library cap.)

Because resolving the sense requires reading your actual sentence, that sentence is treated as private. It is never added to the shared word-explanation cache, and it is sent to a third-party AI (Gemini) only at the moment you tap the in-context explanation — never in the background, never automatically — under a strict no-training policy (your input is never used to train AI models or reused for anything else).

On macOS you can request the in-context explanation right from the capture overlay, on an opt-in button beneath the sentence, before the word is even saved. It is always something you choose, never a wall pushed at you mid-capture, which is what keeps a powerful AI feature on the calm side of the capture flow.

There is a quieter point worth making: by default, your saved sentence already carries the in-context meaning for free. Capecho keeps the exact sentence you met the word in, and seeing the word sitting in real context is often enough to fix the sense without spending anything.

The metered in-context explanation is the opt-in, on-demand version of that same idea — there for the sentences where the construction is genuinely hard and you want the sense spelled out. The free word explanation and your own preserved sentence handle the rest, which is why the daily cap so rarely gets in the way.

Resolving a word in context is not a one-time convenience; it is what makes the later review honest. Because the contextual meaning is saved with the word and your sentence, your spaced-repetition cards test understanding in situ — fronted by the exact sentence, scheduled by FSRS — the way you will actually meet the word again.

That is a sharper test than a bare word-and-definition card, which only ever asks you to recall a definition you never used. And whenever you want them elsewhere, your context-rich cards export cleanly to Anki and CSV.