Skip to main content
Glama
xliff-processor

xliff-processor

XLIFF MCP Server

An MCP (Model Context Protocol) server for processing XLIFF and TMX translation files. This server provides tools for parsing, validating, and manipulating translation files commonly used in localization workflows.

Features

  • XLIFF Processing: Parse and extract translation units from XLIFF files

  • TMX Processing: Parse and extract translation units from TMX files

  • Tag Preservation: Special processing mode that preserves inline tags for AI translation

  • Validation: Validate XLIFF and TMX file formats

  • Translation Replacement: Replace target translations in XLIFF files

  • CSV / JSON Export: Generate CSV or JSON file content from XLIFF and TMX inputs

  • Agent Skills Module: Provide a standalone skills/ module that AI agents can read directly

Related MCP server: @langapi/mcp-server

Installation

python setup.py

Manual Installation

Using pip

pip install -e .

Using the install script

./install.sh  # Unix/Linux/macOS
install.bat   # Windows

Configuration

For Claude Desktop

Add the server to your Claude Desktop configuration file:

macOS/Linux: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %AppData%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "xliff-processor": {
      "command": "python",
      "args": ["-m", "xliff_mcp.server"],
      "cwd": "/absolute/path/to/xliff-mcp-server"
    }
  }
}

Or if using uv:

{
  "mcpServers": {
    "xliff-processor": {
      "command": "uv",
      "args": ["run", "python", "-m", "xliff_mcp.server"],
      "cwd": "/absolute/path/to/xliff-mcp-server"
    }
  }
}

Available Tools

process_xliff

Process XLIFF content and extract translation units.

Parameters:

  • file_name (string): Name of the XLIFF file

  • content (string): XLIFF file content

Returns: JSON with translation units including:

  • fileName, segNumber, unitId, percent, source, target, srcLang, tgtLang

process_xliff_with_tags

Process XLIFF preserving inline tags for AI translation.

Parameters:

  • file_name (string): Name of the XLIFF file

  • content (string): XLIFF file content

Returns: JSON with translation units preserving original formatting tags

validate_xliff

Validate XLIFF content format.

Parameters:

  • content (string): XLIFF file content to validate

Returns: JSON with validation status, message, and unit count

replace_xliff_targets

Replace target translations in XLIFF file.

Parameters:

  • content (string): Original XLIFF file content

  • translations (string): JSON array of translations with segNumber/unitId and aiResult/mtResult

Returns: JSON with updated XLIFF content and replacement count

process_tmx

Process TMX content and extract translation units.

Parameters:

  • file_name (string): Name of the TMX file

  • content (string): TMX file content

Returns: JSON with translation units including metadata

validate_tmx

Validate TMX content format.

Parameters:

  • content (string): TMX file content to validate

Returns: JSON with validation status and unit count

export_xliff_file

Generate CSV or JSON file content from an XLIFF file.

Parameters:

  • file_name (string): Name of the source XLIFF file

  • content (string): XLIFF file content

  • output_format (string): csv or json

  • preserve_tags (boolean): Whether to preserve inline tags before export

Returns: JSON with generated file_name, mime_type, content, and unit_count

export_tmx_file

Generate CSV or JSON file content from a TMX file.

Parameters:

  • file_name (string): Name of the source TMX file

  • content (string): TMX file content

  • output_format (string): csv or json

Returns: JSON with generated file_name, mime_type, content, and unit_count

Available Skills

Skills are separate from the MCP runtime and its tools.

  • MCP Runtime: The xliff_mcp/ package that exposes tools, prompts, and resources

  • Agent Skills Module: The top-level skills/ directory that AI agents can read directly

  • Runtime Workflow Registration: The xliff_mcp/workflows/ package that maps workflow guidance into MCP prompts/resources

The runtime still exposes workflow guidance through:

  • Prompts: Reusable workflow prompts that guide the client through the right tool sequence

  • Resources: A discoverable skill catalog at skills://catalog and per-skill detail resources at skills://{skill_name}

For agent usage inside the repository:

prepare_xliff_for_translation

Validate XLIFF content, extract translation units, and summarize translation readiness.

translate_xliff_with_tags

Extract tag-preserving XLIFF segments and guide AI translation without breaking inline markup.

replace_xliff_targets_from_translations

Merge translated segment JSON back into the original XLIFF and verify the replacement count.

inspect_tmx_translation_memory

Validate TMX content, inspect language pairs, and summarize translation memory entries for reuse.

Usage Examples

Once configured in Claude Desktop, you can use the tools like this:

  1. Process an XLIFF file: "Please process this XLIFF file and show me the translation units"

  2. Validate XLIFF format: "Can you validate if this XLIFF content is properly formatted?"

  3. Replace translations: "Replace the target translations in this XLIFF file with these new translations"

  4. Process TMX file: "Extract all translation units from this TMX file"

  5. Generate a CSV export: "Export this XLIFF file as CSV and give me the file content"

  6. Generate a JSON export: "Export this TMX file as JSON so I can save it locally"

  7. Use a runtime workflow prompt: "Use the translate_xliff_with_tags prompt to help me translate this XLIFF safely"

Development

Running lint

ruff check .

Running tests

python -m pytest

Running the smoke test script

python test_server.py

Running the server directly

python -m xliff_mcp.server

Requirements

  • Python 3.10+

  • mcp >= 1.2.0

  • translate-toolkit >= 3.0.0

  • lxml >= 4.9.0

  • pydantic >= 2.0.0

License

MIT

Support

For issues and questions, please open an issue on the GitHub repository.

Available Tools

9 tools
export_tmx_fileD

Inputs: content (required), file_name (required), api_key, output_format. Returns: result

export_xliff_fileD

Inputs: content (required), file_name (required), api_key, output_format, +1 more. Returns: result

get_server_infoD

Inputs: api_key. Returns: result

process_tmxD

Inputs: content (required), file_name (required), api_key. Returns: result

process_xliffD

Inputs: content (required), file_name (required), api_key. Returns: result

process_xliff_with_tagsD

Inputs: content (required), file_name (required), api_key. Returns: result

replace_xliff_targetsD

Inputs: content (required), translations (required), api_key. Returns: result

validate_tmxD

Inputs: content (required), api_key. Returns: result

validate_xliffD

Inputs: content (required), api_key. Returns: result

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    C2.1/5.0

    Scored across 9 tools

    Disambiguation5/5

    Each tool targets a distinct action (export, process, validate, replace, get info) with specific file types (XLIFF or TMX). The only potential overlap is process_xliff and process_xliff_with_tags, but the latter explicitly handles tags, making them clearly distinguishable.

    Naming Consistency5/5

    All tools follow a consistent verb_noun pattern in snake_case, e.g., export_tmx_file, process_xliff, validate_tmx. The naming is predictable and uniform across the entire set.

    Tool Count5/5

    With 9 tools covering export, process, validate, replace, and server info for both XLIFF and TMX, the count is well-scoped for a file processing server. No tools feel extraneous or missing.

    Completeness4/5

    The tool surface covers the core lifecycle for XLIFF and TMX: export, process, validate, and replace. Minor gaps exist, such as missing import or merge tools, but for a focused XLIFF processor, it is largely complete.

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      B
      quality
      C
      maintenance
      MCP server for managing translation projects in Weblate. With the ability to search, read and write transltaions.
      20
      37 npm
      24
      MIT
    • A
      license
      Not graded
      quality
      F
      maintenance
      An MCP server for internationalization (i18n) tasks, providing tools to translate, move, list, and remove translation keys in JSON files for the Kilo Code extension.
      24
      MIT