Skip to content

Blog

Product updates, guides, and customer stories.

Phrase and Lokalise Changed Their Pricing: What Happened and Your Options in 2026Guides

Phrase and Lokalise Changed Their Pricing: What Happened and Your Options in 2026

Phrase removed its $135/mo Starter plan and now starts business plans at $1,245/mo. Lokalise restructured to processed-word billing and gated its top tiers. What changed, why the market is moving upmarket, and what a small team can do about it.

July 2, 20266 min read
When AI Translations Break: What Actually Fails, and the Graduation PathGuides

When AI Translations Break: What Actually Fails, and the Graduation Path

LLM translation in CI works, until terminology drifts, a plural rule bites, or a native speaker calls the output insulting. What actually breaks in AI-only localization pipelines, why better prompts don't fix it, and how to add scoring, review and delivery without throwing your pipeline away.

July 1, 20266 min read
How Batch Translation Workflows Break Continuous DeploymentEngineering

How Batch Translation Workflows Break Continuous Deployment

A batch translation handoff is a release-blocking step hiding inside your pipeline. Here is exactly how it breaks continuous deployment, and what continuous localization wired into CI/CD looks like instead.

June 25, 20267 min read
The browser can translate your app now. Here's what it can't do.Guides

The browser can translate your app now. Here's what it can't do.

Chrome 138 and Edge 148 ship a free, on-device Translator API that translates any page instantly, offline, and privately. It is genuinely useful, and it is not a translation management system. Here is the honest demo, the governance gap, and a recipe that uses the browser as a sanctioned fallback while your TMS stays the governed core.

June 17, 20269 min read
Is machine-translated content "AI-generated" under the EU AI Act?Guides

Is machine-translated content "AI-generated" under the EU AI Act?

Article 50 of the EU AI Act applies from August 2, 2026, and was not delayed by the Digital Omnibus. The Commission's draft guidance treats AI translations as content that requires marking. What that means for translated websites and apps, and how a documented human-review workflow keeps you on the exemption path under either reading.

June 13, 20269 min read
You don't have to choose: i18next and inlang, togetherProduct

You don't have to choose: i18next and inlang, together

i18next-cli init --inlang scaffolds an inlang project on your existing translation files: Sherlock, Fink and Paraglide work directly on your i18next JSON, Locize stays the hosted TMS on the same files, and nothing migrates.

June 12, 20265 min read
One command from hardcoded strings to a localized appProduct

One command from hardcoded strings to a localized app

npx i18next-cli localize takes a mono-lingual app (often AI-generated) through instrumenting, key extraction, Locize setup, and AI auto-translation in a single run, ending with confidence-scored translations and CDN delivery.

June 11, 20264 min read
Quality Estimation: confidence scores for AI translationsProduct

Quality Estimation: confidence scores for AI translations

Locize now scores every AI translation with an AI confidence estimate, surfaces low-confidence segments in the editor, and can route them straight into your review workflow.

June 10, 20263 min read
Context tools for AI translation: glossary, style guide, memory & screenshotsProduct

Context tools for AI translation: glossary, style guide, memory & screenshots

New read-only Locize MCP tools give your AI agent the project glossary, style guide, translation memory, and screenshot context, so it can translate, review, and lint terminology right in your IDE.

June 9, 20263 min read