Hanzi Tutor

Hanzi Tutor teaches you to read, write and pronounce simplified Chinese — and tells you precisely what you got wrong. It runs entirely on your device: no account, no ads, no analytics, no subscription. WRITE IT, AND SEE WHAT WAS OFF Write a character with your finger, Apple Pencil, trackpad or mouse. Every stroke is graded against real stroke geometry: its shape, where you put it, how much ink you used, and whether it came in the right order. You get a score out of 100 and the reasons — which stroke is misshapen, misplaced, backwards, missing or out of order — in colour on the board. • Trace mode with a faint guide, and Recall mode that shows only the pinyin and meaning • Stroke-order animation that shows the direction of each stroke, not just the order • 7,744 characters in 775 frequency-ordered lessons of ten • Spaced repetition: hard characters come back soon, known ones stop wasting your time SAY IT, AND SEE YOUR TONE Hold the button and say the character or word. Your voice's pitch is drawn as a curve over the shape each tone should have — one chart per syllable, so you see which syllable went wrong and how. The neutral tone is scored too. Words are scored as actually spoken, with tone sandhi applied (你好 is níhǎo), and the app says when it has. If there was not enough voice to judge, it says "not sure" instead of guessing. Tone scoring measures pitch directly. It needs no speech model and no network, and works from the first launch. DRILL THE TONES THAT TRIP YOU UP Minimal tone pairs — 妈 mā, 麻 má, 马 mǎ, 骂 mà — drawn from the app's own dictionary, most common first, with filters for the hard contrasts (2 vs 3, 1 vs 4). Hear a set in order, quiz yourself on which tone you heard, then say each one and see your pitch against the target. ON-DEVICE SPEECH RECOGNITION (OPTIONAL) A tone can be perfect on the wrong syllable: 四 sì and 是 shì can sound alike in pitch. Install the optional recognition model and the app also tells you which syllables it heard — "heard shi where si was asked for" — beside the tone verdict. Recognition runs on your device; your voice is never uploaded. COLOUR CHARACTERS BY TONE Turn on tone colours and every character and its pinyin are drawn in the colour of their tone — on the board, in the word list, in your vocabulary, lessons and radical families. A visual memory aid that ties sound to shape. Off until you switch it on. AND MORE • The full HSK 3.0 word list (9,443 words), searchable by character, pinyin with or without tone marks, or English • Look up any of 9,574 characters: readings, meaning, radical, stroke count, HSK level and every word that uses it • Radicals taught as families, and each character's components, each writable on its own • Your own vocabulary lists with groups — words, phrases, whole sentences — exportable to JSON or CSV • Hear any character or word in your device's own Chinese voice, and graded phrases from bundled recordings PRIVATE AND OFFLINE BY DESIGN The whole course is inside the app: characters, stroke data, word list and audio. Nothing to download on first launch, nothing to sign up for. No analytics, advertising, crash reporting or tracking, and no server of ours. The microphone is on only while you hold the button. What it hears is analysed on the device, never saved and never sent anywhere. The app reaches the network only for three things, each off until you choose it: • An optional speech-recognition model (163 MB), downloaded only when you press the button • An optional speech-synthesis model (about 61 MB) for phrases that have no bundled recording • Optional syncing between your own devices through your own Dropbox account Each download shows its address, size and licence before you start, and is checked against a pinned checksum. Sync sends your study data only to your Dropbox, never to us, and can be protected with Face ID or Touch ID. No account. No ads. No in-app purchases. No streaks to protect.