🔎🀄 LinguaLens 语言透镜 — AI Chinese-learning toolkit: immersive reading, shadowing, translation & fluency coaching. Bring your own key. (React + TS + Vite)
- Category: education product
- Audience: learners, teachers, parents, and education operators who need a clearer learning or exam-prep workflow.
- Repository status: Public repository. Keep examples, docs, and issues free of credentials, private data, and machine-specific paths.
- Topics: ai, chinese-learning, language-learning, learn-chinese, react, tailwindcss, typescript, vite
- 🔎🀄 LinguaLens 语言透镜 — AI Chinese-learning toolkit: immersive reading, shadowing, translation & fluency coaching. Bring your own key. (React + TS + Vite).
- Give users a concrete learning workflow instead of a loose collection of content.
- Make practice, feedback, review, or recommendation steps easier to repeat.
This repository supports productized learning workflows: diagnostic input, guided practice, review loops, and clearer handoff between learner, teacher, and software.
- Primary language: TypeScript
- Detected stack: TypeScript, Node.js, Vite, React, Tailwind CSS
- Default branch:
main - Visibility:
PUBLIC - License: MIT License
srcLICENSEREADME.mdSECURITY.mdindex.htmlpackage-lock.jsonpackage.jsonpostcss.config.jstailwind.config.jstsconfig.jsonvite.config.ts中文学习产品开发技术文档.md
Use the commands that match the current project state:
npm install
npm run dev
npm run previewOn first launch, open Settings and add an AI provider API key:
| Provider | Key format |
|---|---|
| Gemini | Google AI Studio key (AIza...) |
| OpenAI | sk-... |
| Anthropic | sk-ant-... |
Keys are stored locally in your browser and are never sent anywhere except the AI provider you choose. Then pick any feature — paste text into the Reader, snap a photo, or start a Roleplay — and start learning.
Active, single-developer project, built in public. Core features (Reader, Shadowing, Translation, Fluency Coach, SRS, Roleplay, Quests) are usable today; the toolkit grows alongside real classroom feedback from 6,000+ Chinese learners.
Built by Tony (Sheng) — a Chinese-language teacher with 6,000+ students, building AI + Chinese-teaching tools in public.
If LinguaLens is useful to you, ⭐ Star this repo and Follow @shengdabai to follow the journey.
More Chinese-learning tools in the works:
MIT © shengdabai
English | 中文
把 AI 对准任何中文，真正读懂它 —— 然后读出来、说出来、记下来。
语言透镜是一套沉浸式中文学习工具集，用你自己的一把 AI 密钥就能开一间学习工作室：AI 标注阅读、跟读口语练习、双语翻译、口语流利度教练，再加上间隔重复、角色扮演和游戏化任务。自带 Gemini / OpenAI / Anthropic 密钥，全部在浏览器本地运行。
大多数中文 App 把你锁进它们的内容和节奏。语言透镜反过来：素材由你提供 —— 一篇文章、一段聊天记录、一张菜单、你自己写的稿子 —— AI 化身导师，为你标注、讲解、听你的发音，并反复操练直到记牢。无需账号、不在服务器留存你的数据、没有订阅。只要你的密钥和你的素材。
放入中文文本（或随手拍下），语言透镜带你走完完整学习闭环：读懂 → 开口 → 复习 → 记牢。
- 🀄 沉浸式阅读 —— AI 标注中文，即时拼音、释义与划词查询。
- 🎙️ 跟读练习 —— 配合语音识别的跟读，给出发音反馈。
- 🌉 翻译桥 —— 智能双语翻译，并排对照 + TTS 朗读。
- 📖 全能阅读器 —— 多种输入格式的统一阅读体验。
- 🗣️ 流利度教练 —— AI 评估口语流利度，给出可操作的改进建议。
- ✍️ 写作起草 —— AI 辅助的中文写作助手。
- 📸 速拍模式 —— 随手捕捉任意中文，立即开始学习。
- 💬 实时会话 —— 实时对话练习。
- 🔁 间隔重复（SRS） —— 间隔重复系统，长期记忆词汇。
- 🎯 任务系统 —— 游戏化挑战，保持学习连续性。
- 🎭 角色扮演 —— 基于场景的对话练习。
- ⚙️ 个人资料与设置 —— 多服务商 AI 配置（Gemini、OpenAI、Anthropic）。
- 框架： React 18 + TypeScript + Vite
- 样式： Tailwind CSS
- AI 服务商： Google Gemini、OpenAI、Anthropic（用户自配，自带密钥）
- 音频： Web Speech API + TTS
git clone https://github.com/shengdabai/LinguaLens.git
cd LinguaLens
npm install
npm run dev在浏览器打开 **http://localhost:5173**。
构建生产版本：
npm run build| Command | Purpose |
|---|---|
npm install |
Install project dependencies. |
npm run dev |
vite |
npm run preview |
vite preview |
npm run build |
tsc && vite build |
- Keep real credentials out of the repository. Use local environment files, GitHub repository secrets, or the deployment platform secret manager.
- If a
.env.examplefile exists, treat it as documentation only; never commit filled-in.envfiles. - Before publishing screenshots, demos, or client examples, remove private names, internal paths, account IDs, and API endpoints.
- The
Repository Hygieneworkflow is a lightweight guardrail, not a replacement for product-specific tests.
- README describes the user, business outcome, and operating boundary.
- Setup or preview commands are current and do not rely on private machine state.
- No real secrets, private user data, or machine-local state are tracked.
- Screenshots, demos, or sample outputs are safe to share publicly when the repository is public.
- Product-specific tests or smoke checks are documented before production use.
- Tighten the fastest path from clone to useful demo.
- Add project-specific screenshots, sample outputs, or a short walkthrough where useful.
- Promote repeated manual steps into scripts, tests, or documented workflows.
- Keep security, privacy, and licensing boundaries explicit as the project evolves.
Maintained by Tony Sheng. This README is written as a business-facing handoff: it should help a future collaborator, client, or reviewer understand why the repository exists, how to inspect it, and what must be true before it is reused or shipped.
如果语言透镜对你有用，请 ⭐ Star 本仓库 并 关注 @shengdabai，一起见证这段旅程。
更多在做的中文学习工具：
MIT © shengdabai