Tool profile
Lingo.dev: cost, fit and limits
Lingo.dev is an LLM-based localization platform made by Replexica Inc. that translates app and content strings with configurable engines, glossaries and AI review runs, and passes model cost through at no markup, for developers who want to meter spend.
What to know first
Lingo.dev has a Sandbox with no base fee, billed pay as you go at $0.50 per engine per month and $2 per MTok, and a Production tier with a $99 monthly base and 5M tokens a day. Enterprise is custom. The pricing page says LLM cost is passed through with zero markup. The terms state that Lingo.dev does not use user content to train AI models.
Lingo.dev at a glance
| Maker | Replexica Inc., Not stated on the vendor pages |
|---|---|
| Free plan | Yes |
| Entry price | $0 base, pay as you go (Sandbox) |
| API access | No |
| Category | Best AI Localization Software in 2026, rank 8 of 12 |
| Website | lingo.dev |
Who Lingo.dev is best for
Developers who want visible AI cost
Sandbox charges for tokens and engines only, and the page says LLM cost is passed through with zero markup, so spend follows usage.
Teams that care about training promises
Section 5.3 of the terms says Lingo.dev does not use user content to train AI models, and third-party providers are bound by no-training agreements.
Enterprises needing residency and SOC 2
The Enterprise tier names a SOC 2 Type II report, SSO, RBAC, BYOK, EU or US data residency and a 99.99% SLA.
Lingo.dev plans and pricing
| Plan | Price | What it includes |
|---|---|---|
| Sandbox | $0 base; $0.50 per engine per month; $2 per MTok | 100K tokens per day with unlimited engines, seats, glossary, brand voice and rules, and 7-day data retention. |
| Production | $99 per month base; $0.50 per engine; $2 per MTok | 5M tokens per day, email support and 30-day data retention. Pricing for engines and tokens is the same as on Sandbox. |
| Enterprise | Custom | Unlimited throughput, Jira, Zendesk Help Center and Payload CMS integrations, SOC 2 Type II report, SSO, RBAC, BYOK and a 99.99% SLA. |
Checked November 2026.
What Lingo.dev does
- Quality AI review
- Review runs cost $0.01 each on Sandbox and Production, and they flag meaning changes and terminology errors against the glossary before output ships. Reviewers can read those flags before release, which keeps spot checks cheap.
- Engine suggestions
- Engine suggestions and model rankings by language pair use published benchmarks, which gives a team a starting point for choosing a model per locale. Teams that already know their preferred model can ignore the ranking and set an engine directly.
- Enterprise pipeline
- An async pipeline with pre-localization AI edits, post-localization AI review, back-translation checks and selective human review is an Enterprise item. It lets a localization manager decide which strings pass automatically and which need a human.
- Cost reporting
- Glossary coverage and per-model LLM cost reports are Enterprise items that show which model spent the budget on which language. Finance and engineering teams can then see where token spend goes by language and model.
- Brand voice and rules
- Glossary, brand voice and rules are included even on Sandbox, so a free-to-start workspace can still constrain tone and terms. Unlimited engines and seats on Sandbox mean experiments do not need a purchase approval.
How Lingo.dev handles your data
Section 5.3 of the terms states that Lingo.dev does not use User Content to train AI models and that third-party AI providers are bound by no-training agreements. Source code is not stored, third-party AI providers may keep data up to 30 days for abuse monitoring, and users have 30 days to export after termination. The primary database is in the EU (Frankfurt) through Supabase, with application logic on the Cloudflare edge. A SOC 2 Type II report and EU or US data residency choice are Enterprise items. Governing law is Delaware.
Where Lingo.dev falls short in practice
- Country not stated. The pages we read do not state the legal entity's country, only that the terms are governed by Delaware law. Procurement teams may need the registration details from the vendor.
- Best features sit on Enterprise. Human review routing, async pipelines, cost reports, SOC 2 Type II and residency are Enterprise items, so Sandbox and Production users get the engine and glossary only.
- Short retention windows. Sandbox keeps data for 7 days and Production for 30 days, which suits stateless pipelines but not teams that want a long review history inside the tool.
Works with
Alternatives to Lingo.dev
Lingo.dev: common questions
Does Lingo.dev have a free plan?
The pricing page names no free plan. The Sandbox tier has a $0 base and bills pay as you go with credits and auto-reload, at $0.50 per engine per month and $2 per MTok. It gives 100K tokens per day, so a small project may spend little.
What does the Lingo.dev Production plan add?
Production adds a $99 monthly base, 5M tokens per day, email support and 30-day data retention over Sandbox. Engine and token prices stay at $0.50 per engine per month and $2 per MTok. Pipeline features and SOC 2 Type II remain Enterprise items.
Does Lingo.dev train AI on my content?
Section 5.3 of the terms says Lingo.dev does not use User Content to train AI models, and third-party AI providers are bound by no-training agreements. Providers may retain data up to 30 days for abuse monitoring. Ask for the provider list if that retention conflicts with your policy.
Where is Lingo.dev data hosted?
The terms describe a primary database in the EU (Frankfurt) through Supabase and application logic on the Cloudflare edge. EU or US data residency choice is listed as an Enterprise item. Confirm with the vendor whether non-Enterprise workspaces can fix a region.
What is back-translation checking in Lingo.dev?
The pricing page lists a back-translation check among Enterprise review steps, alongside pre-localization AI edits and post-localization AI review. Back-translation converts output back to the source language so a reviewer can spot meaning drift. The pages give no accuracy figures, so test it on your own content.