1. Conduid
  2. Content
  3. Capy
MCP server · Content

Capy

MCP server that inspects a repo, returns a structured /preview brief, and writes a design-system JSON artifact for AI coding agents

37Low

Scored 6 months ago · breakdown

About Capy

Capy is an MCP server in the Content category: mCP server that inspects a repo, returns a structured /preview brief, and writes a design-system JSON artifact for AI coding agents.

Install

Claude Code
claude mcp add capy -- npx -y capy-mcp
npx
npx -y capy-mcp

This server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.

Tools (3)

get_design_systemWrites a machine-readable design-system JSON artifact (`.capy/design-system.json`)

Parameters not published.

get_preview_briefReturns repo map, inspection plan, constraints, and a `/preview` spec for the agent to follow

Parameters not published.

update_previewDiffs against last snapshot, refreshes the design system, returns an incremental update brief

Parameters not published.

Ask AI

✦

Ask AI about Capy

Powered by Claude · Grounded in docs

I know everything about Capy. Ask me about installation, configuration, usage, or troubleshooting.

Security checks

  • ✓README presentThe repository ships documentation.

5 checks have not been run yet.

README


What it does

Capy inspects your React / Next.js repo and returns structured briefs so AI agents can implement UI changes with precision — no hallucinated components, no guessed tokens.

Three tools:

Tool Purpose
get_preview_brief Returns repo map, inspection plan, constraints, and a /preview spec for the agent to follow
get_design_system Writes a machine-readable design-system JSON artifact (.capy/design-system.json)
update_preview Diffs against last snapshot, refreshes the design system, returns an incremental update brief

Quick start

npx capy-mcp@latest

Setup

Pick your client. One command or one config change — you're set.

Agents

claude mcp add --scope user capy -- npx -y capy-mcp@latest

Or add to ~/.claude/mcp.json:

{
  "mcpServers": {
    "capy": {
      "command": "npx",
      "args": ["-y", "capy-mcp@latest"]
    }
  }
}

Add to your config file:

  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
  • Windows: %APPDATA%\Claude\claude_desktop_config.json
{
  "mcpServers": {
    "capy": {
      "command": "npx",
      "args": ["-y", "capy-mcp@latest"]
    }
  }
}
codex mcp add capy -- npx -y capy-mcp@latest

Or add to ~/.codex/config.toml:

[mcp_servers.capy]
command = "npx"
args = ["-y", "capy-mcp@latest"]

Add to ~/.gemini/settings.json:

{
  "mcpServers": {
    "capy": {
      "command": "npx",
      "args": ["-y", "capy-mcp@latest"]
    }
  }
}

Add to ~/.vscode-server/data/User/mcp.json:

{
  "servers": {
    "capy": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "capy-mcp@latest"]
    }
  }
}

Make sure Copilot Chat is in Agent mode — MCP tools only work there.

Add to ~/.vibe/config.toml:

[[mcp_servers]]
name = "capy"
transport = "stdio"
command = "npx"
args = ["-y", "capy-mcp@latest"]
kimi mcp add --transport stdio capy -- npx -y capy-mcp@latest

Or add to ~/.kimi/mcp.json:

{
  "mcpServers": {
    "capy": {
      "command": "npx",
      "args": ["-y", "capy-mcp@latest"]
    }
  }
}

Add to ~/.config/opencode/opencode.json:

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "capy": {
      "type": "local",
      "command": ["npx", "-y", "capy-mcp@latest"],
      "enabled": true
    }
  }
}

Add to settings.json:

{
  "mcpServers": {
    "capy": {
      "command": "npx",
      "args": ["-y", "capy-mcp@latest"]
    }
  }
}

IDEs

Add to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "capy": {
      "command": "npx",
      "args": ["-y", "capy-mcp@latest"]
    }
  }
}

Add to ~/.vscode-server/data/User/mcp.json:

{
  "servers": {
    "capy": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "capy-mcp@latest"]
    }
  }
}

Or via Command Palette: MCP: Open User Configuration

Add to ~/.codeium/windsurf/mcp.json:

{
  "mcpServers": {
    "capy": {
      "command": "npx",
      "args": ["-y", "capy-mcp@latest"]
    }
  }
}

Add to your settings file:

  • macOS: ~/.zed/settings.json
  • Linux: ~/.config/zed/settings.json
  • Windows: %USERPROFILE%\AppData\Roaming\Zed\settings.json
{
  "context_servers": {
    "capy": {
      "command": "npx",
      "args": ["-y", "capy-mcp@latest"],
      "env": {}
    }
  }
}

Add to your config file:

  • macOS: ~/.gemini/antigravity/mcp_config.json
  • Windows: %USERPROFILE%\.gemini\antigravity\mcp_config.json
{
  "mcpServers": {
    "capy": {
      "command": "npx",
      "args": ["-y", "capy-mcp@latest"]
    }
  }
}

Or via UI: Agent session → … dropdown → MCP Servers → Manage MCP Servers → View raw config.


Library usage

import {
  buildPreviewBrief,
  buildDesignSystemArtifact,
  runPreviewUpdate,
  detectFramework,
} from "capy-mcp";

const brief = await buildPreviewBrief(process.cwd(), { task: "build_preview" });
const designSystem = await buildDesignSystemArtifact(process.cwd());
const update = await runPreviewUpdate(process.cwd());

Local development

cd capy-mcp
npm install
npm test
npx capy-mcp

README copy captured when this server was indexed; it may lag the repository. The original is authoritative.

Questions

About Capy

How do I install Capy?

Run claude mcp add capy -- npx -y capy-mcp, then add the server to your MCP client's configuration.

Is Capy safe to use with an AI agent?

Its trust score is 37 out of 100 (low). It passes 1 of the 1 static security checks run so far. It has no ConduID identity yet, so agent calls to it are not receipted.

Is Capy still maintained?

Conduid hasn't recorded a commit date for this repository yet. Check the repository directly for recent activity.