Documentation

Set up GoShip to observe your AI-assisted development workflow.

What is GoShip?

GoShip uses an MCP (Model Context Protocol) server to passively observe how you work with AI coding assistants. It captures behavioral signals — tool calls, edit patterns, iteration cycles — without ever reading your source code. These signals are scored across four dimensions: AI Fluency, Output Quality, Judgment, and Speed.

IDE Setup

Choose your AI coding assistant below:

Claude Code

Run this in your terminal:

claude mcp add goship -e GOSHIP_TOKEN=your-token -- npx -y @goship/mcp-server

Cursor

Add to your project's .cursor/mcp.json:

{
  "mcpServers": {
    "goship": {
      "command": "npx",
      "args": ["-y", "@goship/mcp-server"],
      "env": {
        "GOSHIP_TOKEN": "your-token"
      }
    }
  }
}

Windsurf

Add to your project's .windsurf/mcp.json:

{
  "mcpServers": {
    "goship": {
      "command": "npx",
      "args": ["-y", "@goship/mcp-server"],
      "env": {
        "GOSHIP_TOKEN": "your-token"
      }
    }
  }
}

Privacy

Captured

Tool calls and MCP interactions
File create/modify/delete events
Git commits and diffs
Terminal commands
Prompt/response metadata
Edit frequency and patterns

Never Captured

Source code content
Environment variables / secrets
Keystrokes
Screen recordings
Personal files outside project
API keys or credentials

Scoring Dimensions

25%

AI Fluency

How effectively you communicate with and leverage AI tools.

30%

Output Quality

Code quality, architecture, functionality, and security.

25%

Judgment

Critical evaluation of AI suggestions — accepting, modifying, or overriding.

20%

Speed

Development velocity and time efficiency.

Ready to get started? Create an account for personalized setup instructions.