文档定价研究企业版招聘
招聘中
登录注册预约演示

定价
按用量付费。

LLM 成本零加价直付。基础设施费用 $2/MTok 起。所有套餐均包含完整平台功能。

Mistral AIFreenetEdyoucatedSolanaLaurelCal.comTruelyMistral AIFreenetEdyoucatedSolanaLaurelCal.comTruely
沙盒版
按量付费
适合试用和评估
开始构建
生产版
基础费用 + 使用量费用
适用于生产环境和多租户场景
升级到生产版
企业版
定制
人工审核、合规支持、专属服务
预约演示
使用量
基础价格$0$99/月定制
引擎$0.50/引擎/月$0.50/引擎/月定制
LLM 成本成本直通，0 加价成本直通，0 加价成本直通，0 加价
Lingo.dev 基础设施$2/MTok$2/MTok定制
吞吐量10万 tok/天500万 tok/天无限制
AI 质量审核次数$0.01/次$0.01/次定制
计费方式点数余额，自动充值点数余额，自动充值年度合同
平台
引擎无限制无限制无限制
引擎开通 API是是是
按引擎跟踪用量是是是
席位无限制无限制无限制
术语表无限制无限制无限制
品牌语气无限制无限制无限制
指令无限制无限制无限制
集成GitHub、GitLab、BitbucketGitHub、GitLab、Bitbucket全部功能 + Jira
数据保留7 天30 天定制
API
同步 + 异步 API是是是
管理 API / MCP是是是
Webhooks是是是
治理与质量
可配置的异步流水线是
本地化前 AI 编辑是
人工参与审核是
本地化后人工审核按词计费，0 加价
本地化后 AI 审核是
回译检查是
报告
翻译日志和质量评分是是是
完整治理报告是
按模型划分的 LLM 成本报告是
术语表覆盖率和变更率是
支持
社区是是是
电子邮件是是
专属 Slack是
专属客户经理是
安全与合规
SOC 2 Type II是是是
SSO是
角色与权限（RBAC）是
个人和服务 API 密钥是
BYOK（自带密钥）是
数据驻留（欧盟 / 美国）是
正常运行时间 SLA99.99%
审计日志是
开始构建升级到生产版预约演示
“有了 Lingo.dev，我们的工程师几乎不用再操心本地化。他们只要专注开发功能，翻译就会自动覆盖 36 种语言。”
Keith Williams

Keith Williams

Cal.com 工程负责人

Estimate your monthly cost

LLM cost is passed through at zero markup. Reviews trigger on every translation — their cost scales with translation activity and the reviewers you attach.

Step 1

Choose your plan

Production
$99/mobase

For production and multi-tenant use

  • 30-day retention
  • 5M tok/day throughput
  • Email support
Need SSO, audit logs, custom SLA, or pipeline stages at scale? See Enterprise in the plan comparison above.
Step 2

Configure your engines

Add one engine per translation workflow. Each can have its own model, locales, and pipeline.

~$115.55/ mo
~26K output tok/day
≈ ~65K tokens

Source + glossary terms + custom instructions + brand voice. Default.

Translation
Tokens (~2.3M in · ~780K out)~3.1M
input = 50,000 words × 1.3 tok/word × 10 locales × 3.5× config = ~2.3M · output = 50,000 words × 1.3 tok/word × 10 locales × 1.2× expansion = ~780K
LLM cost · OpenAI: GPT-4o pass-through~$13.49
(~2.3M × $2.5 + ~780K × $10) / 1M
Lingo.dev infrastructure~$1.56
~780K output × $2 / 1M
Quality
AI reviews · $0.01 / run~$100.00
ceil(50,000 words × 10 locales / 500 words per request) × 10 reviewers = 10,000 runs × $0.01 per run
Engine subtotal~$115.55/ mo
Monthly estimate
Total
~$214.55
/ month
Account daily output tokens (all engines)~26K / ~5.0M plan cap
Plan subscription~$99.00
Engines (1 × $0.50)~$0.50
Translation across all engines~$15.05
Quality reviews across all engines~$100.00
How we calculate›
Per engine, per month
source_tokens = words × 1.3volume = source_tokens × locales
Words/month is the actual monthly translation volume — already accounts for re-translations and content updates.
1.3 tok/word — average tokens per English word. Non-Latin scripts push this 2–3× higher:
~1.3
English, Spanish, French, German, Italian, Portuguese
~2.0
Russian, Greek, Hebrew, Arabic
~2.5–3.0
Chinese, Japanese, Korean, Thai, Devanagari (Hindi, etc.)
Translation
input_tokens = volume × config_multiplieroutput_tokens = volume × 1.2llm_cost = (input × model.in + output × model.out) / 1Minfra = output × $2 / 1M  (output tokens only)daily_cap_usage = Σ engines.output / 30  (account-wide, output tokens only)
Config multiplier — every request also carries prompt + glossary + instructions + brand voice on top of the source:
1.5×
Minimal — source only
3.5×
Standard — glossary + instructions (default)
6.0×
Premium — full context
1.2× expansion — translated output is typically 10–30% longer than the source. Per target language:
+20–35%
German
+15–20%
French
+15–25%
Spanish, Portuguese, Italian
+10–30%
Russian
−10–20%
Chinese, Japanese, Korean (often shorter)
1.2× is a global average across typical European target locales.
Calibration source: estimated, pending measurement against engine_request_logs (input_tokens / word_count).
Quality AI reviews
requests = ceil(words × locales / 500 words/request)review_runs = requests × reviewers_per_enginereview_cost = review_runs × $0.01
500 words/request — average source chunk size per translation API call. Lingo.dev splits source content into chunks at sentence / element boundaries; 500 words is the typical chunk for documentation and UI content.
reviewers_per_engine — every AI reviewer (scorer) attached to the engine runs on every translation request. 2 reviewers × 500 requests = 1,000 review runs.
Flat per-run fee — the LLM cost behind each review is absorbed by Lingo.dev, not passed through.
Calibration source: estimated, pending measurement against avg(word_count) in engine_request_logs.
Total
total = plan_base + Σ engines × ($0.50 + translation + reviews)

All numbers above are approximate (~). Estimate assumes typical prompt overhead (glossary, instructions, brand voice). Actual bill varies ±30% with content density and pipeline configuration.

Language script matters: token counts shift significantly across writing systems. Non-Latin scripts (Chinese, Japanese, Korean, Arabic, Hebrew, Thai, Devanagari, etc.) can change the words-to-tokens ratio by 2–3×, and target-language expansion (e.g. English → German typically +20–30%) further shifts the output side.

平台

本地化 API异步任务 API本地化引擎语言检测Lingo.dev Platform MCP定价

开发者工具

Lingo React MCPLingo CLILingo GitHub ActionLingo React Compiler
Alpha

资源

文档Labs指南更新日志语言LLM 模型

公司

博客研究预约演示客户案例招聘
招聘中
humans.txt

社区

GitHubDiscordTwitterLinkedIn
总部位于旧金山，团队遍布全球
SOC 2 Type II·CCPA·GDPR
由 Y Combinator
Combinator
 & Initialized Capital
Initialized Capital
 & 以及我们的客户支持
隐私·条款·Cookies·security.txt

© 2026 Lingo.dev (Replexica, Inc).

所有系统运行正常
登录注册预约演示