superbox.ai

Typical workflows start by installing the SuperBox CLI (pip install superbox). Use superbox init to scaffold a new MCP server project, then superbox push to publish it to the central registry. Clients can be auto-configured with superbox pull --client <client-name>, which writes a correct MCP config pointing to the Lambda endpoint. For interactive experimentation or testing, use superbox run --name <server> to start an interactive session against the Lambda executor, or superbox logs --name <server> to stream execution logs from AWS. The platform also includes search, inspect, and test commands to locate servers, view security reports, and run server tests directly from the repository. The Lambda-backed executor specifically supports Python MCP servers, so Python-based MCPs are the primary target for online execution and sandboxing.

Prerequisites:

• Python 3.8+ and pip installed on your system
• Optional: AWS credentials configured if you plan to use Cloud logs or Lambda execution

Installation steps:

1. Install the SuperBox CLI: pip install superbox

2. Verify installation: superbox --version

3. Authenticate (optional but recommended for registry access): superbox auth login

4. Initialize a new MCP server project (example): superbox init --name my-mcp-server

5. Push your MCP to the registry: superbox push --name my-mcp-server

6. Configure a client to use your MCP via Lambda endpoint: superbox pull --client cursor

7. Run and test locally against the Lambda executor: superbox run --name my-mcp-server

Notes:

• If you are deploying in AWS or using the registry, ensure your environment has appropriate AWS permissions.
• The Lambda executor currently supports Python MCP servers; ensure your MCP is Python-based or compatible with the supported runtime.

Tips and common issues:

• Ensure your MCP server is discoverable by the SuperBox registry (published via push) and includes proper metadata for automated discovery.
• When using logs, you can follow real-time output with superbox logs --name <server>.
• If you encounter registry or authentication issues, run superbox auth status to verify your session, and check your AWS credentials if using Lambda-backed execution.
• The CLI supports client auto-configuration for multiple clients (e.g., Cursor, VS Code, Windsurf, Claude, ChatGPT); use the pull command with the desired client to generate correct MCP config files pointing to the Lambda endpoint.
• For debugging locally, you can use the interactive run mode which opens a session against the Lambda executor; examine logs and results to diagnose environment or runtime mismatches.