Clue2App MCP Server

MCP (Model Context Protocol) server for Clue2App, enabling AI coding assistants like Claude Code to deploy applications directly.

Features

• Seamless Deployment: Deploy apps from within your AI assistant conversation
• Project Management: Create and switch between projects
• App Lifecycle: Create, rebuild, delete applications
• Monitoring: Check build status, view logs, get runtime info
• Custom Domains: Register domains, assign apex (@) and subdomains, manage TLS — all via CLI proxy

Installation

pip install clue2app-mcp

Configuration

Claude Code

Add to your ~/.claude/claude_code_config.json:

{
  "mcpServers": {
    "clue2app": {
      "command": "clue2app-mcp",
      "args": []
    }
  }
}

Claude Desktop

Add to your Claude Desktop config:

{
  "mcpServers": {
    "clue2app": {
      "command": "python",
      "args": ["-m", "clue2app_mcp.server"]
    }
  }
}

Prerequisites

1. Install and authenticate with Clue2App CLI:

   pip install clue2app-cli
   c2a login
   

2. Set up a project context:

   c2a project create MyProject            # namespace is auto-generated
   c2a project use MyProject
   

Available Tools

Auth & Projects

Apps

Custom Domains

Secrets

Daari LLM Gateway

Example Usage

In Claude Code or Claude Desktop:

"Create a new project called 'demo' with namespace 'demo'"

"Deploy my FastAPI app from https://github.com/user/fastapi-app"

"Check the status of my-app"

"Show me the logs for my-app"

"Rebuild my-app"

"Register example.com with DELEGATION method"

"Point example.com (apex) and www.example.com at my-app"

"Show me the NS records I need to add at my registrar"

Development

# Clone the repo
git clone https://github.com/clue2solve/clue2app-mcp.git
cd clue2app-mcp

# Install in development mode
pip install -e ".[dev]"

# Run tests
pytest

Releasing

Releases are cut by pushing a v<version> tag that matches the version in pyproject.toml. The .github/workflows/release.yml workflow then:

1. Builds a wheel + sdist with python -m build.
2. Verifies the tag matches the pyproject.toml [project].version; fails the run if they diverge.
3. Publishes both artifacts to PyPI as clue2app-mcp — only if the PYPI_API_TOKEN repo secret is set. Without the token the publish step is skipped with a clear warning and the build artifacts are still uploaded to the run so the release itself is not blocked.
4. Creates a GitHub Release for the tag with the wheel and sdist attached.

Typical cut:

# 1. Bump the version in pyproject.toml (e.g. 0.1.0 -> 0.1.1) and commit.
git commit -am "chore(release): v0.1.1"

# 2. Tag and push.
git tag v0.1.1
git push origin main --tags

The tag push triggers the workflow; watch it under Actions → Release.

One-time setup (still pending — see SCK-441 follow-ups):

• Claim the clue2app-mcp name on PyPI (requires a human PyPI account).
• Add a PYPI_API_TOKEN repo secret scoped to that project under Settings → Secrets and variables → Actions.

Until both are done the workflow will build and attach artifacts but skip the PyPI upload with a warning.

License

MIT