Talks
Reading alone won't make you speak. Talks are short back-and-forth conversations you listen to — then speaks back, one line at a time, with real voices.
What makes a Talk different
A Talk is a dialogue between two or more speakers — each with its own voice. You hears the conversation spoken aloud, line by line, then practices saying each line back. This is the bridge from recognizing Chinese to producing it.
Each speaker can be assigned a distinct voice in Settings → Talk voices, so a conversation between "Mom" and "Me" actually sounds like two different people.
The app ships with beginner-friendly sample talk categories and conversations — short, everyday dialogues:
- Getting Started — What's Your Name?
- Everyday Talk — Hungry?, Time for School, Who Is This?
Sample speakers are Tom, Mary, Mom, Dad, and Me. Listen to these first; then build your own — and when you want a clean slate for your family's talks, Settings → Data → Clear samples gives you a fresh start.
Create your own talk
-
Make a categoryOn the Talks page, tap 📝 New. Name it by theme — e.g. "Food", "Travel", or "Shopping" — and tap 📝 Create.
-
Start a new talkOpen the category and tap 📝 New. Enter a Title (e.g. "At a restaurant").
-
Pick your speakersUnder Speakers (pick at least 2), tap the people in the conversation — existing names appear as pills, or tap + Add to type a new one like "Waiter". You need at least two. In the New Talk editor, you can type dialogue lines directly — English first, then ✨ Translate to Chinese.
Pick speakers — tap a name pill, or + Add to create your own (at least 2). -
Add the linesTap 📝 Create, then in the talk's Build mode, use Add line: pick a speaker, type or 🎤 speak the English, then ✨ to translate to Chinese, then tap Add. Between existing lines, tap the thin + button to insert a new line right there — no need to delete and re-add the lines below.
-
Edit any line in placeMade a typo or want to reword a line? Tap the ✎ pencil at the end of any line in Build mode — the line opens as a small editor right where it is: pick a different speaker, retype or 🎤 speak the English, tap ✨ to refresh the Chinese, then 💾 Save. Nothing else moves.
A talk rebuilt from a real past conversation — with a waiter, a teacher, your mom, a friend from last night — is remembered far better than a made-up textbook dialogue. The Chinese lines attach to real people and real moments, so review feels like replaying a memory, not memorizing. Pick your speakers, say each line in English, tap ✨ — your past becomes your practice.
-
Or let AI draft itIn an empty talk, scroll below Add line to find ✨ Generate with AI (optional). Describe a scene (e.g. "ordering food at a restaurant") and tap Generate. The AI writes 4–6 bilingual lines using your chosen speakers, which you can then edit.
Picked the wrong speakers, or want to add a new character partway through? On the talk list, tap the ✎ edit button on any talk. The same speaker picker appears with the current speakers already selected — add, remove, or create new ones, then save. The talk opens right back up so you can keep writing lines.
The magic is in speaking back. After listening to a line, say it out loud before moving on — mimicking the speaker's voice and tone. Production, not just recognition, is what builds fluency.
Listening and speaking a talk — Immerse mode
Open any talk and toggle to 📖 Immerse to read it as a clean, distraction-free conversation. The same ▶ Play all, ⏱ Timer, and Speed controls from Stories work here too — and each line keeps its own ▶ button so you can replay any speaker. Toggle back to ✏️ Build to add, edit, or remove lines.
Focus on sounds alone — hide the text
Immerse mode has four toggles in its toolbar. Turning them off does not silence the line — the ▶ play button always stays, so you can hear the sound and recall the characters themselves:
| Toggle | What it shows / hides |
|---|---|
| Pinyin ā | The pinyin above the text. When a tone changes in connected speech (tone sandhi), a second row appears showing how it's actually said — with the changed syllables in red. |
| Translate 汉→EN | The whole-line English translation. |
| Line U | The underline beneath each clickable word. |
| Hide 字 | The Chinese characters themselves (the layout stays, but the glyphs vanish). |
🔊 Sounds & tones — how many sounds does Chinese have?
About 400 syllables — that's Mandarin's entire sound inventory. Add the 4 tones (plus 1 neutral) and you get about 1,300 sounds. Every word in the language is covered by that set.
Same syllable, four tones, four words — tap to hear the difference:
- 妈 mā (mom)
- 麻 má (hemp)
- 马 mǎ (horse)
- 骂 mà (scold)
Wrong tone, wrong word — which is why tone combinations matter in every sentence. Words flow together and tones interact: two 3rd tones in a row flip the first one (你好: written nǐ hǎo, said ní hǎo — see tone sandhi below).
Compare with English: English pitch only carries feeling, and each word's pronunciation is memorized one by one — an adult English speaker has effectively memorized the pronunciation of roughly 20,000 words. Chinese reuses ~400 building blocks (~1,300 with tones) — the tone is what separates them. English even has a mini version of tones: stress — REcord (noun) vs reCORD (verb) changes the word. You don't learn hundreds of tones; you learn 4, then train your ear to catch them everywhere. Master that, and you can pronounce any Chinese word you meet.
🎵 Tone sandhi — two pinyin rows
A tone isn't always spoken the way it's written. When two 3rd tones meet, the first becomes a 2nd — 你好 is written nǐ hǎo but said ní hǎo. With Pinyin on, the app shows two rows so you sees the change:
- Green — the dictionary pinyin. Blue — how it's actually said (what the ▶ voice speaks).
- The blue row appears only when a tone shifts, and the exact syllables that changed are drawn red (for 我想买水果: wó and shuí).
A sample talk shows it on every line — here it is on a phone:
This is automatic in both Build and Immerse modes. It applies to Stories too.
🧩 English vs Chinese — why the words don't line up
After you tap ✨ and your English becomes Chinese, the two lines won't match word-for-word — Chinese grammar is not harder, just different. One big pattern — Chinese tells the sentence like a story, in the order things happen — plus three small surprises cover most of what a beginner meets.
- Chinese tells the sentence like a story. English decorates a sentence — the, a, to, will. Chinese mostly drops the decorations: 我去商店 (wǒ qù shāngdiàn) — “I go store”: who, what, where, already complete. Add time, and time comes first: 今天五点,我去商店 (jīntiān wǔ diǎn, wǒ qù shāngdiàn) — “Today 5 o'clock, I go store.” Two events? Tell them in order: 我和妈妈先去商店,然后吃饭 (wǒ hé māma xiān qù shāngdiàn, ránhòu chīfàn) — “I and Mama first go store, then eat.”
- Little words sit in the middle. 也 yě (also) and 都 dōu (all) have a fixed seat: right after who, right before what. Say 妈妈「也」累了 (māma yě lèi le) — “Mama is also tired” — not ❌ 也妈妈累了 (yě māma lèi le). Pattern: who → also → what.
- No “is” before describing words. The word 饿 è already means “is hungry”, so 是 shì isn't needed: say 我很饿 (wǒ hěn è), not ❌ 我是饿 (wǒ shì è). That 很 hěn just keeps the sentence sounding natural — it doesn't mean “very” here.
- The question word stays home. 你吃面 (nǐ chī miàn) — “You eat noodles” → hide the noodles → 你吃「什么」 (nǐ chī shénme) — “You eat what?” 谁 shéi (who) and 哪里 nǎlǐ (where) work the same way.
And you never have to study these rules: learn 妈妈「也」累了 (māma yě lèi le) as one phrase — your own mama, your own memory — and the correct word order comes inside it, free. That is exactly why Stories and Talks keep whole sentences rather than single words: each phrase you save is a grammar lesson you never had to take.
Tapping a word — ▶ Play and 📖 Pleco
Tap any word (even a hidden one) to open a small dictionary card showing its hanzi, pinyin, and meaning. Two buttons sit inside it:
Pleco is a third-party app and a trademark of Pleco Software Inc.
- 📖 Pleco — opens the word in the Pleco dictionary app (a free, leading Chinese dictionary) for fuller definitions and examples. If Pleco isn't installed, it offers to get it from the Play Store.
- ▶ Play — speaks just that word aloud using the Chinese voice, so you can hear its exact pronunciation.
So for pure listening practice, you can turn on Hide and turn off Pinyin and Translate — leaving only the speaker's name and the ▶ button on each line. They tap ▶, hear the line spoken, and try to recall the characters from sound alone. It's a powerful way to train the ear before the eye. (Tapping where a hidden word sits still opens this dictionary card — with its own ▶ and Pleco — if they get stuck.)
Like stories, the words in your talks automatically become flashcards — so every conversation you build also feeds their review deck.
Tap the stars under any talk to rate it 1–5. Use the sort dropdown to order by Rating, or tap the ☆ / ★ / ★★ filter chips to show only your favorites. Category ratings are the average of everything inside — so your best talks rise to the top automatically.
In Immerse mode, the 🎤 My recordings section lets you record yourself reading the conversation aloud — in character! Tap 🔴 Record, read the dialogue, then tap Save. Your last 3 recordings are kept. Play them back with ▶ to hear how you sound.
After each recording, tap A / B / C / D to grade your own pronunciation. An emoji face appears so you can track progress at a glance.
💡 The app does not automatically check your pronunciation. It works fully offline — no AI listening. Compare your recording to the ▶ TTS audio and be your own judge.
Works exactly like in Stories — see the screenshots in Record yourself reading.
Reading — graded lessons that grow with you →
Flashcards — how review adapts →
Remember — the memory palace →