Weblate

You deploy a Weblate MCP Server to connect an AI assistant with your Weblate translation management platform. It provides full access to Weblate’s API, enabling natural language driven translation workflows, project and component management, and translation operations through a lightweight MCP interface that can run locally or be accessed over HTTP transports.

How to use

You interact with the Weblate MCP Server using an MCP client. The server exposes tools that let you list projects, inspect components, search translations using Weblate’s native filters, read and write translations, and view detailed statistics and history. You can drive common translation tasks with natural language prompts, for example asking to list all projects, fetch translations for a language, update a translation, or create a new project. If you want to run automation, you can also integrate these capabilities into your CI/CD or localization pipelines.

How to install

Prerequisites you need before starting:

• Node.js 18+

• pnpm package manager

• A Weblate instance with API access

Follow these steps to set up and run the MCP server locally or in development mode.

# Clone the repository and install dependencies (example paths shown; adapt as needed)
git clone <this-repo>
cd weblate-mcp
pnpm install

# Configure environment
cp .env.example .env
# Edit .env with your Weblate API URL and token then save

# Build and start for production or local testing
pnpm build
pnpm start

Server runs by default at http://localhost:3001. You can override the port by setting PORT in your environment before starting.

Environment configuration

Set these environment variables to point the MCP server at your Weblate instance and to configure runtime behavior.

WEBLATE_API_URL=https://your-weblate-instance.com/api
WEBLATE_API_TOKEN=your-weblate-api-token
PORT=3001
NODE_ENV=production
LOG_LEVEL=info

Available tools

list_projects

List all available Weblate projects with URLs and metadata

list_components

List components in a specific project with source language details

search_units_with_filters

Efficient search using Weblate's native filtering syntax to find translations based on state, source text, component, and more

search_string_in_project

Search for translations containing specific text within a project

get_translation_for_key

Get translation value for a specific key in a given project and component

write_translation

Update or write translations with optional approval flow

bulk_write_translations

Batch update multiple translations with error handling and efficiency

find_translations_for_key

Find all translations for a specific key across languages

list_languages

List languages available in a specific project

get_project_statistics

Retrieve comprehensive statistics for a project, including completion and string counts

get_component_statistics

Get detailed statistics for a specific component

get_project_dashboard

Provide a complete dashboard overview with all component statistics

get_translation_statistics

Statistics for a specific translation, project, component, and language

get_component_language_progress

Show translation progress for all languages in a component with progress indicators

get_language_statistics

Statistics for a language across all projects

get_user_statistics

User contribution statistics and activity metrics

list_recent_changes

Recent changes across all projects with user and timestamp filters

get_project_changes

Recent changes for a specific project

get_component_changes

Recent changes for a specific component

get_changes_by_user

Recent changes by a specific user