Deploy applications from AI coding assistants like Claude Code
MCP Endpoint: https://clue2app-mcp.control.apps.clue2.app/mcp/sse
This MCP (Model Context Protocol) server enables AI coding assistants like Claude Code and Claude Desktop to deploy applications directly to the Clue2App platform without leaving the conversation.
Supports both local stdio (for Claude Code) and remote SSE/HTTP (for cloud-hosted assistants).
| Tool | Description |
|---|---|
c2a_auth_status | Check if authenticated with Clue2App |
c2a_list_projects | List all projects the user has access to |
c2a_get_current_project | Get the currently selected project context |
c2a_use_project | Set the active project context by name or ID |
c2a_create_project | Create a new project with name and namespace |
c2a_list_apps | List all applications in the current project |
c2a_get_app | Get detailed information about an application |
c2a_create_app | Deploy a new application from a Git repository |
c2a_rebuild_app | Trigger a rebuild of an application |
c2a_delete_app | Delete an application |
Use the hosted MCP server (no installation needed):
{
"mcpServers": {
"clue2app": {
"url": "https://clue2app-mcp.control.apps.clue2.app/mcp/sse"
}
}
}
Note: Remote mode requires setting C2A_ACCESS_TOKEN environment variable or authenticating first.
Run the MCP server locally (uses local ~/.c2a credentials):
{
"mcpServers": {
"clue2app": {
"command": "uvx",
"args": ["clue2app-mcp"]
}
}
}
Before using the MCP server, authenticate with the CLI:
pip install clue2app-cli c2a login
See the CLI Quick Start for full setup instructions.
Once configured, you can ask Claude:
# Deploy an app "Create a new app called 'my-api' from https://github.com/user/repo" # Check status "What's the build status of my-api?" # View logs "Show me the logs for my-api" # Manage projects "List all my projects" "Switch to the 'production' project"