VSCode Translate

A VS Code extension that translates Markdown files and saves the result as a side-by-side _zh.md file, inspired by kiss-translator.

Open any .md file, run the command, and a translated version opens next to the original — translations appear in real-time as each batch completes.

Features

File-based output — generates xxx_zh.md alongside the original, open both side-by-side
Translation cache — saves to xxx.translate.json, skips unchanged paragraphs on re-run
Incremental display — see translations fill in block by block, no waiting for the full file
Works with any OpenAI-compatible API — Zhipu (GLM), DeepSeek, OpenAI, Ollama, etc.
Smart markdown protection — image captions are translated, URLs/paths/code are preserved

Quick Start

1. Configure API

Open VS Code settings (Ctrl+,) and set:

Setting	Description	Example
`vscodeTranslate.apiEndpoint`	OpenAI-compatible base URL	`https://open.bigmodel.cn/api/paas/v4`
`vscodeTranslate.apiKey`	Your API key	`your-api-key`
`vscodeTranslate.model`	Model name	`glm-4-flash`, `deepseek-chat`
`vscodeTranslate.targetLanguage`	Target language	`Chinese` (default)

2. Translate

Open a .md file in VS Code
Click the zh button in the editor title bar, or press Ctrl+Shift+P → run "Translate Markdown (Bilingual Preview)"
The translated xxx_zh.md opens on the right, translations appear as they complete
Run the command again on the same file — cached blocks load instantly, no API cost

3. File layout

my-paper/
├── my-paper.md              # Original
├── my-paper_zh.md           # Generated translation
└── my-paper.translate.json  # Translation cache

Supported API Providers

Any service that provides an OpenAI-compatible /chat/completions endpoint:

Provider	Endpoint	Model
Zhipu (GLM)	`https://open.bigmodel.cn/api/paas/v4`	`glm-4-flash`
DeepSeek	`https://api.deepseek.com/v1`	`deepseek-chat`
OpenAI	`https://api.openai.com/v1`	`gpt-4o-mini`
Ollama (local)	`http://localhost:11434/v1`	`llama3`

Architecture

src/
├── extension.ts         # Command registration and activation
├── fileTranslator.ts    # Main workflow: parse → cache → translate → write
├── markdownParser.ts    # Split markdown into text/code segments
├── translator.ts        # OpenAI-compatible API client with URL protection
└── translationCache.ts  # JSON-based cache keyed by text hash

Markdown is split into segments by blank lines; code blocks are skipped
Image/link URLs are replaced with placeholders before translation, restored after — so ![caption](https://github.com/gulai22/VSCodeTranslate/raw/HEAD/path) gets the caption translated but the path preserved
Consecutive text segments are batched (~3000 chars) to reduce API calls
After each batch, the editor content updates in-place via TextEditor.edit()

Development

npm install           # Install dependencies
npm run compile       # Production build (esbuild → dist/extension.js)
npm run watch         # Dev build with watch + sourcemaps
npm run lint          # TypeScript type check (tsc --noEmit)

Press F5 in VS Code to launch Extension Development Host for testing.

License

MIT

Markdown AI Translate

Gulai