DocsPricingResearchEnterpriseCareers
Hiring
Sign inSign upBook a Demo

Pricing
Pay for what you use.

LLM costs passed through at zero markup. Starting from $2/MTok for infrastructure. Full platform on every tier.

Mistral AIFreenetEdyoucatedSolanaLaurelCal.comTruelyMistral AIFreenetEdyoucatedSolanaLaurelCal.comTruely
Sandbox
Pay as you go
For tinkering and evaluation
Start Building
Production
Base + usage
For production and multi-tenant use
Upgrade to Production
Enterprise
Custom
Human review, compliance, dedicated support
Book a Demo
Usage
Base price$0$99/moCustom
Engines$0.50/engine/mo$0.50/engine/moCustom
LLM costPass-through, 0 markupPass-through, 0 markupPass-through, 0 markup
Lingo.dev infrastructure$2/MTok$2/MTokCustom
Throughput100K tok/day5M tok/dayUnlimited
Quality AI review runs$0.01/run$0.01/runCustom
BillingCredits, auto-reloadCredits, auto-reloadAnnual contract
Platform
EnginesUnlimitedUnlimitedUnlimited
Engine provisioning API
Per-engine consumption tracking
SeatsUnlimitedUnlimitedUnlimited
GlossaryUnlimitedUnlimitedUnlimited
Brand voiceUnlimitedUnlimitedUnlimited
InstructionsUnlimitedUnlimitedUnlimited
IntegrationsGitHub, GitLab, BitbucketGitHub, GitLab, BitbucketAll + Jira
Data retention7 days30 daysCustom
APIs
Sync + Async API
Management API / MCP
Webhooks
Governance & quality
Configurable async pipeline
Pre-localization AI edit
Human-in-the-loop review
Post-localization human reviewPer word, 0 markup
Post-localization AI review
Back-translation check
Reports
Translation logs and quality scores
Full governance reports
Per-model LLM cost reports
Glossary coverage and change rates
Support
Community
Email
Dedicated Slack
Dedicated account manager
Security & compliance
SOC 2 Type II
SSO
Roles & permissions (RBAC)
Personal & service API keys
BYOK (bring your own keys)
Data residency (EU / US)
Uptime SLA99.99%
Audit logs
Start BuildingUpgrade to ProductionBook a Demo
“Now with Lingo.dev, our engineers don't even think about localization. They just build features, and translations happen automatically in 36 languages.”
Keith Williams

Keith Williams

Head of Engineering at Cal.com

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.

Platform

Localization APIAsync Jobs APILocalization EnginesLanguage DetectionLingo.dev Platform MCPPricing

Developer Tools

Lingo React MCPLingo CLILingo GitHub ActionLingo React Compiler
Alpha

Resources

DocumentationLabsGuidesChangelogLanguagesLLM Models

Company

BlogResearchBook a DemoCustomersCareers
Hiring
humans.txt

Community

GitHubDiscordTwitterLinkedIn
HQed in San Francisco + worldwide
SOC 2 Type II·CCPA·GDPR
Backed byY Combinator
Combinator
&Initialized Capital
Initialized Capital
&our customers
Privacy·Terms·Cookies·security.txt

© 2026 Lingo.dev (Replexica, Inc).

All systems normal
Sign inSign upBook a Demo