The Problem with Key-Based Translation
For years, developers have relied on key-based translation (i18n) for their applications. You extract strings, assign them keys like home.welcome_message, and send them off for translation.
This works perfectly for UI buttons, labels, and short messages.
But what about long-form content?
When you need to translate:
- Documentation
- Blog posts (like this one!)
- Legal agreements
- Marketing landing pages
Key-based translation breaks down. You end up with hundreds of fragmented keys (doc.paragraph_1, doc.paragraph_2), completely stripping away the context. Translators can't see the flow of the document, resulting in disjointed and awkward translations.
Introducing Document Translation
Today, we're excited to announce Document Translation for IntlPull—a new way to localize your long-form content without losing context.
How It Works
Instead of extracting keys manually, you simply upload your Markdown (.md, .mdx) files directly to IntlPull. Our intelligent parser breaks the document down into logical blocks (headers, paragraphs, lists) while preserving the document structure.
The Zen Editor
We've built a brand new Split-Pane Editor specifically for documents.
- Left Pane: Your original source document.
- Right Pane: Your translation in real-time.
- Sync Scrolling: As you scroll the source, the target pane follows, so you never lose your place.
Smart Features
- Markdown Parsing: We handle all the syntax for you. Bold, italics, links, and code blocks are preserved automatically.
- Variable Ghosting: Use variables like
{{userName}}in your specialized content? We automatically "ghost" (lock) these tokens so translators can't accidentally break your code. - Frontmatter Support: We automatically extract and translate YAML frontmatter (like title and description) while keeping the keys intact.
CLI Integration
We know developers love the command line. That's why Document Translation is fully supported in the IntlPull CLI.
Terminal1# Upload a document 2intlpull documents upload README.md --source en --target de,fr 3 4# List documents 5intlpull documents list 6 7# Download translated file 8intlpull documents download --id <doc_id> --language de --output README.de.md
Automate with the MCP Server
For those building AI agents, we've also added document capabilities to our Model Context Protocol (MCP) server. Your AI agents can now list, read, and translate documents programmatically.
Availability
Document Translation is available starting today for all customers on the Growth plan and above.
