Development · MODEL CONTEXT PROTOCOL
Lingo.dev
⚡️ Open-source AI-powered CLI for web & mobile localization. Bring your own LLM or use Lingo.dev localization engine. Join discord: https://lingo.dev/go/discord
连接能力
Lingo.dev 为 AI Agent 提供什么
来源目录未提供结构化功能列表,请在使用前阅读源仓库文档。
安装目标
软件包或命令参考
lingo-dev该字段仅作为目录提示,不是已验证的安装命令;请以当前源仓库为准。
安全
MCP 收录不等于安全审核
MCP 服务器可能获得模型上下文、凭据、本地文件或调用外部系统的权限。连接 Agent 前,请检查代码、环境变量、网络行为、软件包来源与维护状态。
README
仓库文档
Lingo.dev is the localization engineering platform: the best way to measure translation quality, translate with LLMs, and proofread with native speakers.
A localization engine is a stateful translation API your team configures and Lingo.dev runs. Build one per product, per content type, or per brand. Every request through an engine applies everything you configured in it, in a fixed order of precedence:
| Layer | What you configure | From the docs |
|---|---|---|
| LLM models | Which model handles each language pair, with ranked fallbacks | 400+ models; the response names the model that ran |
| Brand voice | How your product speaks in each language, one text per locale | Tone and formality per market |
| Rules | The linguistic conventions a generic model misses | Adjective position in Spanish, a space before percentage signs |
| Glossary | Exact term mappings per locale, matched by meaning | "911" becomes "112" for European markets; product names pass through |
| AI reviewers | Scoring that runs after every translation | GEMBA scores, BERTScore, glossary compliance |
Glossaries, rulesets, and brand voices belong to your organization, and an engine applies them by attachment. One glossary governs five engines, and one edit reaches all five. Test a change in the Playground before it goes live: compare an engine against a raw model, or two engines side by side. The engines are configured on the platform, where the localization team runs the localization infrastructure.
Translate the content in a repository. lingo push sends the files to the engine named in .lingo/config.json, and lingo pull writes the translations back from any machine:
npm install -g @lingo.dev/cli
lingo init && lingo link
lingo pushOr call an engine directly, naming it by ID:
const res = await fetch("https://api.lingo.dev/process/localize", {
method: "POST",
headers: { "X-API-Key": process.env.LINGO_API_KEY, "Content-Type": "application/json" },
body: JSON.stringify({
engineId: "eng_abc123",
sourceLocale: "en",
targetLocale: "de",
data: { greeting: "Hello, world!", cta: "Get started" },
}),
});
const { data, model, usage } = await res.json();
// data: { greeting: "Hallo, Welt!", cta: "Jetzt starten" }
// model: "anthropic/claude-sonnet-4.5"
// usage: { inputTokens: 2789, outputTokens: 861, cost: 0.023012 }| Lingo.dev MCP | Your coding agent creates an engine, adds glossary terms, tunes rules, and compares two engines, from the conversation where the problem surfaced |
| Lingo.dev CLI | Push source files, pull translations, from a terminal or from CI. Eighteen formats: JSON, YAML, Markdown, MDX, PO, XLIFF, Flutter ARB, Android and Xcode strings, SubRip, PHP |
| Lingo.dev in CI/CD | Install the CLI and run lingo push as a step in GitHub Actions, GitLab CI/CD, Bitbucket Pipelines, or any runner with Node.js 22+ |
| Lingo.dev GitHub App | Install once and every push to the default branch opens or updates a translation pull request, or translations land as a commit in the pull request that changed the source. No runner, no API key secret, no lockfile to manage |
| Lingo.dev API | One synchronous call per language pair, or an async job that fans one request out to many locales and delivers results as they land |
社区