MCP for
AI engineers

Connect your agent to SimpleLocalize and let it manage keys, translations, and publishing through the Model Context Protocol. Works with Cursor, Windsurf, GitHub Copilot, Claude Desktop, and custom agent runtimes.

Start for free
No credit card required•14-day free trial•Tracking-free service
Configuring SimpleLocalize MCP server in Cursor AI mcp.json settings

Why teams use it

Your agent can talk to localization data

Translation management becomes part of the same workflow where code is written and reviewed.

Stay inside your editor

Ask your AI agent to add or update translation keys directly from source code context.

Batch translation updates

Run bulk updates for missing translations across languages with one prompt.

Publish from agent

Trigger hosting publication for a selected environment after your review pass.

Watch the workflow

MCP localization in Cursor AI

See a real session where an AI agent uses the SimpleLocalize MCP server to work with translation keys and translations directly from the editor.

Installation

MCP config in 30 seconds

Create a project API key in SimpleLocalize and add this config to your MCP client. Replace SIMPLELOCALIZE_API_KEY with your key value.

mcp.json
{ "servers": { "simplelocalize": { "type": "stdio", "command": "npx", "args": ["-y", "@simplelocalize/simplelocalize-mcp", "--api-key=SIMPLELOCALIZE_API_KEY"] } } }

Capabilities

Exposed MCP Tools

  • createTranslationKeys: Bulk key creation from code
  • updateTranslationsBulk: Multi-language text updates
  • getTranslations: Fetch missing translations by locale
  • publishHosting: Trigger zero-downtime hosting deployment
  • manageLanguages: Add or delete project locales
Agent session output
# Prompt in Cursor > Add missing checkout translations for de and fr # Tool calls getAllTranslationKeys -> 152 keys getTranslations -> 19 missing in de, 22 missing in fr updateTranslationsBulk -> 41 translations updated # Result ✓ de and fr checkout flows are complete ✓ Ready for review before publishing

Local development

Build and debug your own MCP workflow

You can run the server locally, attach MCP Inspector, and test tool responses before sharing configuration with your team.

Terminal
# Install and build $ pnpm install $ pnpm build # Run inspector $ pnpm inspector --api-key=slk_xxxxxxxxx ✓ Inspector running at http://localhost:6274 ✓ Tools discovered: 14 ✓ Ready to test requests and responses

See full setup in MCP documentation and source code on GitHub.

CTA FeaturesGo to Editor

Connect your AI agent to localization

  • Cursor, Windsurf, Copilot, Claude Desktop support
  • Bulk key and translation operations
  • Hosting publish from MCP tools
Create free project
No credit card required∙5-minute setup
"The product
and support
are fantastic."
Laars Buur|CTO
"The support is
blazing fast,
thank you Jakub!"
Stefan|Developer
"Interface that
makes any dev
feel at home!"
Dario De Cianni|CTO
"Excellent app,
saves my time
and money"
Dmitry Melnik|Developer

Frequently asked questions

Which AI clients support SimpleLocalize MCP?

Any client supporting standard Model Context Protocol (stdio transport) works out of the box, including Cursor, Windsurf, Claude Desktop, and GitHub Copilot.

Do I need a paid SimpleLocalize account?

No, the MCP server works with free-tier SimpleLocalize projects using your project API key.

How does the agent know which translation keys to create or update?

The agent reads your codebase context (such as React useTranslation hooks, i18next calls, or JSON translation files) and compares key usages directly against your SimpleLocalize project using the getAllTranslationKeys tool.

Can I use the SimpleLocalize MCP server in CI/CD pipelines or headless agents?

Yes. The MCP server runs via npx or Node.js over standard input/output (stdio), making it compatible with custom agent runtimes, CLI workflows, and automated scripts.