开口兽
Practice everyday English, one useful sentence at a time.
Sign in and connect to load original scenario practice from the cloud. Type, deliberately paste or use on-device speech input to look up words and translate English to Chinese or Chinese to English. Save what you want to revisit.
Preparing for launch · Not publicly available on the App Store yet
The first planned release is for iPhone and iPad, requiring iOS / iPadOS 18.6 or later. The learning interface is primarily Chinese, with original English/Chinese scenarios; full Japanese or Korean interfaces are not offered. The Mac app is in development with no confirmed release date.
Cloud scenarios and check-in progress
Sign in to fetch the original catalog over HTTPS with stable unique IDs and a consistent revision. Choose bilingual dialogues, blind listening or shadowing by category. Scenarios stay only in current-run memory, not a bundled or downloaded offline bank; initial loading and restarting require a connection. Deliberate check-ins and undo save completed scene IDs with your account to restore progress and count completed or untried scenarios, not assess ability.
One input, two translation directions
Type or tap Paste. English-to-Chinese is the default: English words show IPA, common senses and bilingual examples, and sentences show a Chinese translation. Switch to Chinese-to-English for English translations of Chinese words or sentences with a short Chinese explanation. Authorized queries run after an input pause; AI results are references, not an authoritative dictionary.
Optional on-device speech input
Tap the blue voice button to request microphone and system speech-recognition permission, then recognize English or Chinese on device for the selected direction. Unsupported device/language combinations report an error without cloud-recognition fallback. Audio and transcripts stay in memory; final text is queried only with separate cloud permission.
On-device English speech
The complete Kokoro model ships with the app. Each request synthesizes fresh audio and plays it from memory without uploading speech text or saving generated audio. Initial model loading takes time; voices are synthetic.
Personal collections
Sign in with Apple to save words and sentences and restore them across devices. Name and avatar are optional, and account deletion is available in Settings. Cloud queries and collection sync require a connection.
Understand what leaves your device
Typed text, deliberately pasted text and final on-device recognition text require separate permission before being sent to Microsoft Azure OpenAI. After permission, queries run automatically after an input pause; permission can be withdrawn in Settings. Both translation directions share usage limits; our server does not retain source text or translations, while generic English dictionary entries are cached separately. Provider data and safety terms still apply. System recognition is strictly on-device, without app audio storage or audio upload; local recognition does not itself authorize cloud text queries. AI-query permission is separate from catalog loading: declining AI queries does not disable local English playback of loaded content, but fetching or reloading scenarios still requires sign-in and a connection. There is no pronunciation scoring, advertising or subscription.
Help and contact
Provided by Hangzhou Xulian Information Technology Co., Ltd. Contact info@xunova.tech for product and privacy questions.
info@xunova.techIs it a good fit for you?
You learn primarily through Chinese and want everyday English scenarios, listening and reading alongside word and sentence lookup.
Before you get started
Preparing for iPhone and iPad launch, with iOS or iPadOS 18.6+ planned. No public download yet. Scenario loading and collection sync need sign-in and a connection.
Before you download
Practical answers about what the app does, and what it does not promise.
How does 开口兽 support everyday English practice?
Choose original English and Chinese dialogues by scenario and explore blind listening, Chinese-assisted reading and shadowing. Type, deliberately paste or use optional local speech input for word and sentence queries, then save what you want to review. Check-ins record progress, not language ability.
Does on-device speech input mean offline translation?
No. Optional speech input recognizes on device and reports an error without cloud-recognition fallback when unsupported. AI queries for the final text still need separate permission and a connection. Initial scenario loading and restarting also require sign-in and a connection; this is not a preinstalled offline bank.
Is it available now, and does it score pronunciation?
The iPhone and iPad app is still preparing for launch, without a public download. It does not score speaking or pronunciation, or turn check-ins into an ability certificate. The learning interface is primarily Chinese.