Skip to content

Speaking and Listening

github-actions[bot] edited this page Oct 22, 2026 · 1 revision

Speaking and listening

Voices

Fluenough uses your phone's own text-to-speech. It adds nothing to the download and needs no network (ADR-0002). A word has a speaker button on its card: tap it to hear the word.

Under Settings > Activities and memory, "Play words automatically" plays a word as a card shows, or once you answer. You can also long-press a speaker to turn it on or off. If sound is off, questions that need sound are skipped, and a deck's page offers to turn sound on.

Voices for each language

Settings > Activities and memory > Voices is a page of its own. It lists each language you learn and says whether this phone has a voice for it. A language without a voice skips listening. To add one, open your phone's Settings, search for "Text-to-speech", open the preferred engine's settings and install the voice data. Fluenough installs nothing itself. Voices from other apps are listed under "Other voices"; install one, come back, and the app checks again.

When a language's own voice is missing, a backup voice may read the card's other spelling. The page says so. Where a language has more than one voice, you can choose one. Play speaks a word of the course in the voice in use, at your speech rate. "Say something" opens a speaking test. It tries the phone's speech recogniser, not a voice: you say a course word or anything, and it shows what the phone heard.

Listening drills

You hear a word, then choose, and then type, what it means. A word can name a minimal-pair partner, a word that sounds close to it, and listening offers it among the meanings. The separate drill on sound contrasts is built as a screen but not switched on yet, so no lesson asks it.

Speaking drills

You see the meaning and say the word. The phone's speech recogniser listens.

  • Microphone. Fluenough asks to use the microphone. If it is refused, speaking stays off. If Android will not ask again, allow it in your phone's settings, under Apps, Fluenough, Permissions.
  • Wrong or almost right. Try again lets you say it again as often as you like. Hear yourself plays your voice back, then the model's. Only your first answer is scheduled.
  • "Didn't catch that." Say it again. If the phone has no recogniser for the language, speaking for it stays off.
  • Online recognition. If your phone cannot recognise a language by itself, the app asks whether to recognise it online. What you say then goes to your phone's speech service. It is per language and you can switch it off.

A few languages' speech cannot be recognised on a phone yet. The deck says so.

Offline

Speech and voices work offline when the voice data is on the phone. Recognition that needs the online service does not.

See also: Learner guide, Home.

Clone this wiki locally