i18nstack

Translation backlogs kill launch velocity. i18n Agent targets that specific bottleneck: it accepts source localization files in nine formats — JSON, YAML, CSV, XML, Markdown, Properties, and others — runs them through a multi-model AI pipeline, and returns translated files with original structure preserved. The core workflow is file-in, file-out, with quality scores attached to each translation batch. An API is available for teams that want to wire this into CI pipelines rather than triggering it manually.

The differentiating feature is the MCP client integration. Rather than requiring a separate dashboard context switch, the vendor ships a Model Context Protocol client installable with a single npx command into Claude Code, Cursor, and VS Code. Translations happen in-editor, with the tool handling API key management and caching automatically. For developers already working in AI-assisted editors, this collapses the localization step into the same tool loop they use for everything else.

This tool fits teams shipping SaaS products who treat localization as a recurring engineering task and need to scale to 50+ languages without proportional headcount growth. The vendor cites project costs in the €150–300 range for work that traditional agencies price at €15,000+, and claims a 99% cost reduction to approximately €0.01 per word. Where the ceiling appears is in workflows that require a human translator to review, approve, or dispute specific strings before they reach end users — the tool has no built-in review stage, approval workflow, or translator seat model. Teams with those requirements layer a separate review tool on top, which means managing two systems.

Self-hosting is not available, so teams with data residency requirements or air-gapped environments cannot run this on their own infrastructure. The open-source MCP client code is available to inspect, but the translation service itself runs on the vendor’s infrastructure. GDPR compliance and SSL encryption are listed on the page; teams in regulated industries should verify whether the vendor’s infrastructure meets their specific data handling requirements before routing production content through the service.