Guides / VS Code with GitHub Copilot

Candor UI with VS Code with GitHub Copilot

Put the docs MCP server in .vscode/mcp.json (the key is servers, not mcpServers), install the skill, and build a chat.

  1. 01Add the registry

    In a React 19 app with Tailwind 4 (Next.js or Vite), init adds the @candor registry to components.json, installs the theme and imports it after Tailwind; --yes answers shadcn's questions, so a coding agent can run it. Without the CLI, add the registry to components.json by hand and install with npx shadcn@latest add @candor/<name>.

    Set up Candor
    Add the chat core
    JSON · components.json
    {
      "registries": {
        "@candor": "https://candor.themethodic.com/r/{name}.json"
      }
    }

    See:Get started

  2. 02Add the docs MCP server

    The docs MCP server at https://candor.themethodic.com/mcp is remote (streamable HTTP), read-only and needs no key: it searches these docs, reads any page and suggests components for a task. It serves docs, never paid source. VS Code reads .vscode/mcp.json in the workspace, or the mcp.json in your user profile. Its top-level key is servers, not mcpServers. MCP: Add Server in the Command Palette does the same through a form: choose HTTP and paste the address. code --add-mcp takes the same entry with a name and adds it to your user profile.

    Your user profile, from the terminal
    JSON · .vscode/mcp.json
    {
      "servers": {
        "candor": {
          "type": "http",
          "url": "https://candor.themethodic.com/mcp"
        }
      }
    }

    See:VS Code: MCP configuration

  3. 03Install the skill

    The skill teaches which component fits which job and the rules the kit keeps. The skills CLI installs it for this tool into .agents/skills/ in your project (-g for every project).

    Install the skill

    See:The skillllms.txt

  4. 04Try it

    With the server and the skill in place, the agent can look up Thread, Message, Composer and Tool call in these docs and install them with the CLI. Check what it installed against candor-lock.json.

    An example prompt

    See:ThreadComposerTool callPatterns