Quizzz API

English

日本語
繁體中文

English

日本語
繁體中文

Appearance

Listening Questions ​

In short: listening isn't a separate quiz type with its own editor — it's three fixed audio question formats you dial into the normal AI generation flow, either as a dedicated all-listening quiz or added on top of any other AI quiz. Quizzz writes the dialogue, synthesizes two-voice audio for it, and grades it like any multiple-choice question.

Try it ​

This is a real listening question (Quizzz Official's TOEIC Part 1-style demo) — play the audio and pick the matching picture description:

Open in a new tab ↗

Create one ​

There are two ways to get listening questions into a quiz — both go through AI generation, since there's no way to hand-write synthesized dialogue from a blank question.

A dedicated listening quiz ​

  1. Open Quizzz ↗ → New Quiz and pick the Listening quiz — AI questions with audio tile.
  2. Pick your target language and level. The language picker only offers languages Quizzz has text-to-speech voices for (see Language coverage below) — everything else is hidden here.
  3. Choose how many questions of each listening format to generate:
    • Quick response — hear a short utterance, pick the most natural reply.
    • Task comprehension — hear a short A/B dialogue, decide what action should happen next.
    • Key point — hear a short dialogue or explanation, extract one key fact (time, place, reason, item).
  4. Describe the listening scenario you want (e.g. "station announcements and asking for directions" or "ordering at a restaurant") — the same free text box the topic-based AI track uses, just prompted for listening material instead of reading material.
  5. Quizzz drafts the dialogue for every question, synthesizes the audio, and opens the result in the normal quiz builder for you to review, edit, and publish — an AI draft is never auto-published.

Added on top of any other AI quiz ​

The same 3-format picker also appears at the bottom of the question mix when you generate a quiz from a topic or from your material — pick a count for any of the three formats and Quizzz generates those as extra questions on top of whatever else you're generating (nothing in your existing mix gets converted to audio). It only appears once your target language has listening support, and it doesn't apply to the Manual or Template tracks, since neither one calls the AI.

Every plan caps a quiz at 20 listening questions total, and Free is additionally capped lower — see Plan limits below.

Hand-edit a generated question ​

Once AI has generated a listening question, open it in the content step to edit it: the on-screen prompt, the dialogue itself line by line (who's speaking — A, B, or a narrator — what they say, and the pause before each line), and each speaker's voice gender where the language offers more than one voice. Nothing re-synthesizes on every keystroke — press Generate audio (or Update audio after an edit) when you're ready, so editing text doesn't quietly spend synthesis calls.

There's no way to start a listening question from a blank line, though — the + Add Question picker in the content step doesn't offer it as a type. Hand-editing only works on a question the AI already produced.

What students see ​

A listening question shows only the format label, an audio player, and the answer options — the spoken script itself is never shown or sent to the browser as text, so a student can't just read the transcript instead of listening. Students can play, pause, seek, and replay the audio as many times as they like (it isn't play-once), and there's a Slow toggle that gently slows the narration without distorting the pitch — useful for lower-level students working through a dialogue.

Listening plays in Classic and Interactive answer modes. Assign Guided (the one-question-at-a-time mode) to a quiz that contains any listening question and Quizzz automatically falls back to Classic — Guided doesn't support audio-only items yet.

Good to know ​

  • Language coverage — listening currently covers 39 languages (Google Cloud text-to-speech voices), including Japanese, Chinese (Taiwanese Mandarin by default, with a Mainland Putonghua accent available for HSK-style content), English (US and UK voices), Korean, Vietnamese, Spanish, French, German, Italian, Portuguese, Arabic, Russian, Hindi, Turkish, Dutch, Indonesian, Polish, and over 20 more. Most languages have both a female and a male voice, so a two-person dialogue sounds like two people; a newer batch of languages currently ships only one Google voice, and the builder tells you when that's the case — every speaker in the dialogue is read by that single voice.
  • Plan limits — Free allows up to 10 listening questions per quiz with up to 60 seconds of audio per question; Lite, Go, and Pro all allow the full 20-question cap, with 120, 180, and 300 seconds of audio per question respectively.
  • AI quota — like any AI-generated content, listening questions draw on your plan's monthly AI token quota. Editing an already-generated question's dialogue afterward is free; only pressing Generate/Update audio spends quota again.

Next ​

TechTrans Lab