Smartling MCP. Manage global translation projects from chat.

Look, you're managing localization jobs in Smartling; you don't wanna be clicking through a bunch of different dashboards just to get strings translated. You need your AI client—your agent—to handle the whole workflow right here. This server lets you talk to Smartling using natural language commands and manage every part of the content process.

Viewing and Inspecting Projects:
You can start by getting an overview of everything you're working on. Use list_projects to see all the translation initiatives managed in your Smartling account. If you need the deep details on just one job, use get_project with a specific project ID; it pulls back all the status info you need.

Preparing Source Content:
Before a translator can do squat, they gotta have content. You send raw source files—think JSON, XML, or whatever resource bundle format you're using—with upload_file. It dumps that stuff right into the system for processing. But wait, it gets better. Translators need context to get it right; they can't just guess what a button says.

You attach visual assets, like screenshots or HTML snippets, using upload_context. This helps your linguists write accurate localized strings because they see exactly how the text appears in the app.

Managing Jobs:
When you're ready to kick off the translation work, you use create_job to set up and initialize a brand new job inside an existing project. Once that job is live, and maybe you realize you missed some strings or need to add more specific content chunks—like legal disclaimers or product names—you don't restart everything.

You just call add_content_to_job, which increments the running job with those extra segments. That keeps your whole process super tight for the linguists.

Getting the Final Output:
The last thing you wanna do is manually download files from a website. When the translations are done, you use download_translated_file. You specify the locale and the content set, and boom—it retrieves the final translated file output right into your environment. It's that simple.

Basically, it handles everything: finding what projects exist (list_projects), digging deep on a specific project (get_project), feeding the raw files (upload_file), giving translators pictures to look at (upload_context), kicking off the work cycle (create_job), adding more strings mid-stream (add_content_to_job), and finally, grabbing the finished goods (download_translated_file). It keeps your whole localization pipeline running without you ever leaving your chat interface.