More than 3,900 Mandarin dialogues are open here, graded HSK 1 through 6 with a 7+ band above that, and listening to them needs no account. Over 800 sit at HSK 1 alone, so beginners are not stuck replaying the same four recordings.
Every dialogue carries audio plus three layers of text — characters, tone-marked pinyin, and English — and each layer has its own switch. That makes it reading practice as much as listening practice, and lets you choose which skill you train. Hide pinyin and English to read hanzi against the audio. Hide the characters to work with your ears only. Turn everything on when a line has beaten you twice and you want to see why. Below the player the full dialogue is laid out as a transcript, and tapping any word opens its dictionary entry — one more tap saves it to spaced-repetition flashcards with a free profile.
Learners who read Chinese well and understand almost nothing spoken are not imagining the gap. Written Chinese hands you the characters one at a time; speech hands you an unbroken stream and expects you to find the seams yourself. Mandarin also runs on a small inventory of about four hundred syllables before tone is counted, so a great many words sound alike and context, not sound, tells them apart. That is why the same sentence can be obvious on the page and impenetrable in the ear, and why more replaying rarely fixes it on its own.
Play the dialogue once at full speed with the text hidden and take only the gist. Replay at 0.75x or 0.5x — the audio genuinely slows rather than dropping in pitch — and repeat single lines until the syllables separate. When a line still refuses to resolve, read it in the transcript before playing it a sixth time: more often than not the missing piece is a seam rather than a word, and once you can see where the words break you will hear it. If a particular sound rather than a boundary is what defeats you, the pinyin chart plays that syllable on its own in the tones it actually occurs in. Finish by shadowing: play a line, pause, and say it back at the same speed and rhythm.
Comprehension lags reading, so start one level below where you read comfortably; a dialogue you follow at roughly nine words in ten is the productive zone. Topics run from news, travel, and sport to film, health, and viral trends, and if none of it interests you, generate a dialogue on your own topic at your level instead. When you can follow a dialogue without the text, read it out loud in speaking practice or browse the other modes on the practice hub.
Yes, and it does not need an account. You can open any dialogue, play the audio, change the speed, and read the transcript while signed out. A free profile only matters if you want to save words to flashcards or keep track of which dialogues you have finished.
Yes. Each dialogue shows characters, tone-marked pinyin, and English alongside the audio, and each layer can be switched off independently. Reading the transcript after a listening pass is the intended workflow, so one dialogue serves both skills rather than forcing you to choose.
Playback runs at 0.5x, 0.75x, or full speed, and you can replay a single line as many times as you need instead of restarting the dialogue. Slower speeds are useful for hearing where one word ends and the next begins, which is the usual reason a line sounds like noise.
Start one level below the level you read comfortably, because listening comprehension develops later than reading. A good fit is a dialogue where you follow roughly nine words in ten — enough to stay in the story while a few unknown words push you forward.
You can generate a new dialogue on any topic you name, at the HSK level you choose, and it comes back with audio, characters, pinyin, and an English translation like every other dialogue in the catalog. Interest matters more than difficulty for staying with listening practice.