This feature is in beta.
This guide shows you how to connect the Lokalise MCP Server to your tools or AI agents and start using it right away. The setup is simple, and you only need your Lokalise API token to get started.
What is the Lokalise MCP Server?
The Lokalise MCP Server lets MCP-compatible AI tools connect to Lokalise and perform supported actions using natural language.
MCP (Model Context Protocol) is an open standard that allows AI assistants to work with external tools and services in a consistent way. With Lokalise MCP Server, you can connect tools such as Claude, Cursor, Windsurf, or other MCP-compatible clients to your Lokalise workspace.
Instead of switching to the Lokalise UI or writing custom API scripts, you can ask your AI assistant to complete supported Lokalise actions for you.
For example, you can ask it to:
check translation progress for a project
list untranslated keys
create a localization task
add target languages
retrieve key details
create or update glossary terms
work with screenshots for localization context
The Lokalise MCP Server is cloud-hosted and supports different toolkits depending on your workflow:
Project management toolkit — for project, task, contributor, language, file, key, and glossary workflows.
Software development toolkit — for developer-focused workflows, such as working with projects, tasks, keys, and screenshots from MCP-compatible development tools.
You can use the MCP Server to reduce context switching, automate common localization actions, and bring Lokalise workflows into the AI tools your team already uses.
How to set up the Lokalise MCP Server
Different MCP-compatible tools (AI agents, IDEs, orchestration platforms, etc.) have their own way of adding MCP servers, but the setup process is usually the same everywhere:
Find the place where the tool lets you add an MCP server or edit MCP configuration.
Create a new MCP server entry pointing to a Lokalise endpoint.
Add your Lokalise API token under the
apikeyheader.You can generate a new Lokalise API token in your Profile settings > API tokens.
Save the configuration so your tool/agent can detect the capabilities.
Below is the standard configuration you can use in any MCP-compatible environment.
Configurations
The Lokalise MCP Server provides two separate toolkits: Project management and Software development. You may need one or both, depending on which actions you want your agent to perform. Some actions require the PM toolkit, others require the SD toolkit.
A full table showing which action belongs to which toolkit is available below.
Project management toolkit configuration
With OAuth 2 (recommended approach):
"lokalise_pm": {
"url": "https://mcp.lokalise.com/mcp/project-management"
}
With API key:
"lokalise_pm": {
"type": "http",
"url": "https://mcp.lokalise.com/mcp/project-management",
"headers": {
"apikey": "[Enter-your-API-key-here]",
"accept": "application/json, text/event-stream"
}
}
Replace [Enter-your-API-key-here] with your actual Lokalise API Key.
Alternatively, you can provide API key as Authorization: Bearer %apikey%.
Software development toolkit configuration
With OAuth 2 (recommended approach):
"lokalise_sd": {
"url": "https://mcp.lokalise.com/mcp/software-development"
}
With API key:
"lokalise_sd": {
"type": "http",
"url": "https://mcp.lokalise.com/mcp/software-development",
"headers": {
"apikey": "[Enter-your-API-key-here]",
"accept": "application/json, text/event-stream"
}
}
Example: Setting up the MCP server in Cursor
Open Cursor Settings.
Go to Tools & MCP.
Find the MCP configuration (JSON file) section.
Add the configurations to your MCP JSON file.
Using the MCP server
After you save the configuration:
Open a new conversation with an agent.
Ask it to use the Lokalise MCP server.
The agent will automatically detect the available Lokalise tools and perform actions in Lokalise on your behalf.
No extra commands or setup needed; the agent understands everything from the MCP configuration.
Frequently asked questions
What is the MCP Server and why use it?
The Lokalise MCP Server lets MCP-compatible AI tools interact with Lokalise in a structured way.
Use it when you want to run supported Lokalise actions from an AI assistant or automation tool instead of switching to the Lokalise UI or writing custom API scripts. For example, you can ask an AI assistant to check project progress, list untranslated keys, create a task, add languages, or update glossary terms.
Who can use the Lokalise MCP server?
Any Lokalise team that’s on a plan with API access can use the MCP Server. If your team’s plan includes API access, you’re good to go!
What kind of API token do I need?
We recommend using OAuth 2.0 to connect the Lokalise MCP Server. OAuth 2.0 is the safer option because you do not need to manually create, copy, or store an API token in your MCP client configuration.
If you still want to use an API token, the token must match the actions you want the MCP Server to perform.
For read-only actions, such as listing projects or checking task progress, a read-only API token is enough.
For actions that create, update, or upload anything in Lokalise, such as creating projects, uploading files, adding tasks, updating keys, or creating glossary terms, you need a read/write API token.
Token permissions follow the user
API tokens in Lokalise always work in the context of the user who created them. It means:
If the user doesn’t have access to a project, the token won’t have access either.
If the user is not allowed to perform an action (like uploading files), the token also won’t be allowed.
If the user has limited project permissions, the token will inherit those limits.
So before using the MCP Server, make sure the user who generates the token has the right permissions for the tasks you want to automate. Learn more about permissions in the Contributors article.
Are there any limits?
Yes — the MCP Server uses the Lokalise API under the hood. This means all API rate limits and restrictions also apply to MCP.
What actions can I complete with the Lokalise MCP Server?
The Lokalise MCP Server currently provides two toolkits:
Project management
Software development
Different actions rely on different toolkits, so make sure you’ve enabled the right configuration(s) from the section above. These two toolkits cover most common automation tasks, and more (including translator-focused ones) will be added over time.
You don’t need to tell your agent which toolkit or action to use; it will automatically pick the correct one based on your prompt.
Projects
Description | Use case | Toolkit |
List all projects | View all projects in the team, including their current status |
|
Create a new project | Create a new localization project with its initial configuration |
|
Get specific project details | Retrieve details for a specific project |
|
Get project key count and statistics | Check project size and progress using key counts and statistics |
|
Languages
Description | Use case | Toolkit |
List project languages | View the base and target languages configured for a project |
|
Add target languages | Add one or more target languages to a project |
|
Contributors
Description | Use case | Toolkit |
List project contributors | View users who have access to a project |
|
Add project contributors | Add users to a project so they can work on localization tasks |
|
Tasks
Description | Use case | Toolkit |
List all tasks | View tasks in a project |
|
Get task details and progress | Check task details, assignees, due dates, and progress |
|
Update task due date and assignees | Change task scheduling or assignment details |
|
Create a task | Create a new localization task |
|
Keys
Description | Use case | Toolkit |
List all keys | View keys in a project, optionally filtered by untranslated status, platform, filename, tags, or key name |
|
Get specific key details | Retrieve details for a specific key, including its translations where available |
|
Add one or multiple keys | Create one or more new keys in a project |
|
Update key details | Update editable key details |
|
Files
Description | Use case | Toolkit |
Upload a file | Upload localization files to a project |
|
Glossary
Description | Use case | Toolkit |
List glossary terms | View glossary terms available in a project |
|
Create glossary terms | Add new approved terminology to the glossary |
|
Update glossary terms | Update existing glossary terminology |
|
Get specific glossary term details | Retrieve details for a specific glossary term |
|
Screenshots
Description | Use case | Toolkit |
Create screenshots | Add screenshots to provide visual context for localization |
|
List screenshots | View screenshots linked to a project |
|
Retrieve a screenshot | Get details for a specific screenshot |
|
Update a screenshot | Update screenshot details or associations |
|
Delete a screenshot | Remove a screenshot from a project |
|

