Settings — Complete Reference

Settings opens locked. Tap Edit to change anything.

Quick jump: Appearance · Reading · Writing · AI · Voice · Talk voices · Recording · Review · Data · Reset

👀 Appearance

How much Chinese you see, and how large the text is.

📖 Reading display

What appears on each line while reading stories and talks. These are the most-toggled settings day to day.

🎵 Tone sandhi — why there are sometimes two pinyin rows

A syllable's tone isn't always spoken the way it's written in the dictionary. When two 3rd tones sit together, the first is actually said as a 2nd tone — 你好 is written nǐ hǎo but spoken ní hǎo. With the Pinyin toggle on, the app shows both so you see the rule in action:

• Green row — the dictionary pinyin (the "look it up" form).
• Blue row — how it's actually said, which is exactly what the ▶ voice speaks. This row appears only when a tone shifts; lines with no change show a single green row.
• Red syllables — in the blue row, the specific syllables that changed are drawn red, so you can spot them at a glance (for 我想买水果, that's wó and shuí).

This works automatically in both Build and Immerse modes, in Stories, Talks, and Reading alike — controlled by the Pinyin display modes:

1 · Both rows — dictionary (green) over spoken (blue) on sandhi words; one row on pure lines.

2 · Spoken only — every word's spoken form; changed syllables stay red.

3 · Dictionary only — the written dictionary tones, one row everywhere.

4 · Smart single row — spoken (red) where tones change, dictionary elsewhere. Exactly one row per word.

✍️ Writing practice

How the stroke-tracing practice behaves. These are great to tune to your age and patience.

Writing practice + Voice settings.

🧠 Learning — AI assistant

The ✨ AI Translate and ✨ Generate features. Entirely optional — without any provider, the ✨ buttons appear but are disabled with a hint to add a key, and everything works by hand.

AI assistant — quality preferences and providers.

Pick a provider and paste an API key — free tiers like Groq or Gemini work.

🤖 No AI key? Use any chatbot

No key, no problem — any free chatbot (ChatGPT, Gemini, Groq…) can do the same job as the ✨ button. Copy a prompt below, paste your story or conversation after it, send, then copy the lines back into Bamboo Island one by one: each CN line goes into the Chinese field, its EN line into the English field (the two always come in pairs). The app then adds pinyin, tap-to-look-up words, and review cards automatically.

The story prompt

Translate my story into simple Mandarin Chinese for a beginner learner.

Rules:
- Use only common, simple characters — the easiest vocabulary a first-year learner knows (HSK 1-2 if you can).
- Short sentences: around 10 characters or less each.
- Simplified Chinese. No pinyin.
- Keep my exact meaning — this is my own memory. Do not add, remove, or change details.
- Write in first person (我) and do NOT invent character names (no 小明 etc.) — only mention a person if my story names them.
- Keep names in English letters (Mary stays Mary, not 玛丽).
- Output one sentence at a time, in this exact format, so I can copy them one by one:

EN: (the English sentence)
CN: (the Chinese translation)

My story:
(paste your story here)

The conversation prompt

Turn our conversation into simple Mandarin Chinese for a beginner learner.

Rules:
- Common, simple characters only (HSK 1-2 if you can). Short lines.
- Simplified Chinese. No pinyin.
- Keep our exact meaning — these are our own words. Do not change details.
- Keep names in English letters (Mary stays Mary).
- Output line by line, alternating speakers, in this exact format:

SPEAKER: (name)
EN: (the English line)
CN: (the Chinese line)

Our conversation:
(paste your conversation here)

HSK level — explained

HSK (Hanyu Shuiping Kaoshi) is the standard scale for Mandarin proficiency, with 6 levels — HSK 1 is the simplest everyday words, HSK 6 is advanced. Setting the level here does two useful things:

• Guides the AI's word choice — when the AI translates or generates Chinese, it prefers vocabulary at or below your level (so a beginner gets simple, common words). It's a soft preference: the AI still uses the correct word when a simple one won't do, so the meaning is never wrong.
• Drives the reading color-coding — when "HSK colors" is on (in Reading display), each word is colored relative to your level: green = known (at or below their level), yellow = learning (one level up), red = advanced (two or more levels up). So as you grow, raise the level and more words turn green.

The default is HSK 1 (best for beginners). As you progress, move it up. "No cap" turns off both effects — the AI uses any word, and everything reads as "known" (no yellow/red).

"View words" — browse the HSK 3.0 vocabulary

Tap 📋 View words next to the HSK level to open the vocabulary browser. It shows the full official HSK 3.0 word list — every word with its hanzi, pinyin, and English meaning — and you can switch between all six levels and search within any level.

The HSK vocabulary browser — every official word, searchable, across all six levels.

Drill into any word — level, pinyin, and meaning at a glance.

Each word card also shows a count (★N) of how many of your own stories and talks use that word — so you can see which level's vocabulary you are actually meeting in practice. Tap a card to jump to where the word appears.

🔊 Voice

The single most important setup for listening practice. The app uses your device's text-to-speech voices — you need a Chinese voice installed for Chinese playback.

Voice — pick the Chinese voice and adjust speed, pitch, and volume.

Talk voices — a different voice for every speaker.

💬 Talk voices

So a conversation sounds like real people talking, each speaker can have its own voice. Here's how it works:

• When you create a Talk and add a speaker (e.g. "Mom", "Dad", "Me"), that speaker automatically appears in this list.
• Assign each speaker a different voice — this is what makes a dialogue sound like two or more distinct people, not the same voice reading every line.
• If you don't assign one, the speaker uses the global default voice from the Voice section above. So assigning voices is optional, but recommended for a natural-sounding conversation.

Talk voices — assign a distinct voice to each speaker.

Audition any speaker with ▶ — or 🔊 Test speakers to hear them all together.

🎤 Recording

Rules for your own read-aloud takes (Immerse mode). This section is independent of the Voice section — it works even when no Chinese text-to-speech voice is installed.

Recording — limit, takes kept, and the microphone status row.

More recording options below — adjust to taste.

🎴 Review (SRS)

How the flashcard deck schedules reviews. The defaults are well-chosen; change only if sessions feel too long or too sparse.

Review — batch size, the daily reminder, and the scheduling controls.

The daily reminder — pick a time, and review becomes routine.

💾 Data — Backup & samples

Your safety net: back up your work, restore it, and manage the sample content.

Backup / Restore — save all data to a file, or restore from one.

Clear / Restore samples — manage only the seeded sample content.

↺ Reset to defaults

If anything feels broken or you want a clean slate for the settings, the ↺ Reset button (in the editing lockbar) returns all preferences to their defaults. It shows a confirmation dialog first: "All settings will return to their defaults. Your AI keys, stories, talks, and cards are kept."

Navigating Settings

The header has two pages of nav — tap the ◀ / ▸ chevron to move between them. Quick-jump chips at the top jump straight to each section (👀 Appearance, 📖 Reading, ✍️ Writing, 🧠 Learning, 💾 Data). Voice and Talk-voices are surfaced via the search bar. Tap Edit to unlock, make changes, then Done.