mcpField
V
vibie-mcp

vibie-mcp

Local
github.com/shdomi8599Cloud & Infra1Updated 22w ago

About vibie-mcp

vibie-mcp lets Claude Desktop, Cursor, and other MCP clients publish static folders to Vibie URLs and update existing sites. OAuth sign-in and a folder marker support repeat deployments.

Key features

  1. 01Create a site from a folder
  2. 02Update an existing site
  3. 03List account sites
  4. 04Retrieve site metadata
  5. 05Skip hidden files and node_modules

Use cases

  1. 01Publish a static HTML folder
  2. 02Redeploy changes to the same URL

MCP support

It is an MCP server that exposes Vibie site creation, update, listing, and metadata tools to compatible clients.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"shdomi8599-vibie": {
"command": "npx",
"args": [
"-y",
"vibie-mcp"
]
}
}
}
Run commandnpx -y vibie-mcp

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npmvibie-mcp0.2.2

Tools

4

Saved tool list · updated 2w ago

ParameterTypeRequiredDescription
namestringNoDisplay name. Defaults to folder name.
notestringNoOptional deployment note (changelog-style).
folderstringYesAbsolute or relative path to the folder containing index.html (or a single HTML file). Required.
categorystringNoCategory id (portfolio / wedding / landing / event / game / photo / linkbio / misc). Defaults to misc.
is_publicbooleanNoWhether the site appears in the public gallery. Defaults to true.

Details

From the project README

vibie-mcp

MCP server for Vibie — deploy static folders to permanent *.vibie.page URLs from Claude Desktop / Cursor / any MCP-compatible client.

Install

One-line auto setup (recommended)
npx vibie-mcp setup

Auto-detects Claude Desktop (Windows Store + APPDATA / macOS / Linux) and Cursor (~/.cursor/mcp.json) configs, then adds the vibie entry. Restart the client and you're done.

Manual config

If auto setup doesn't fit your client, add this to your MCP config (claude_desktop_config.json, ~/.cursor/mcp.json, etc.):

{
  "mcpServers": {
    "vibie": {
      "command": "npx",
      "args": ["-y", "vibie-mcp"]
    }
  }
}

After saving, fully quit and restart your client.

First-time auth

On first tool call, the server initiates an OAuth 2.0 Device Authorization Grant. Your AI will receive instructions like:

Please open https://vibie.io/device?code=XXXX-YYYY in a browser, sign in with Google, and click Authorize. Then ask me to try again.

After authorizing in the browser, ask your AI to retry the same request. The server stores a token in ~/.vibie/credentials.json (chmod 0600) and reuses it for future calls.

You can revoke the token anytime at https://vibie.io/settings/api.

Tools

  • vibie_create_site — Upload a folder and create a new Vibie site. Auto-writes .vibie/site.json in the folder so future updates use the same site.
  • vibie_update_site — Re-deploy to an existing site. Reads slug from .vibie/site.json if not specified.
  • vibie_list_sites — List sites under your account.
  • vibie_get_site — Metadata for one site by slug.

How AI typically uses it

You: "Deploy this folder to vibie"
AI: → vibie_create_site({ folder: "." })
    → Returns: https://my-folder-x7f2.vibie.page

You: "Push my changes"
AI: Detects .vibie/site.json
    → vibie_update_site({ folder: "." })
    → Same URL, new content

Folder structure

What gets uploaded from a folder:

  • index.html + style.css + js/, assets/, etc — all included
  • Hidden files (.git, .vibie, .DS_Store) — automatically skipped
  • node_modules/ — skipped
  • Limits: 100 MB per site, 500 files per upload, 25 MB per file (Vibie's server validates)

A single HTML file (any name) also works — it gets auto-renamed to index.html on upload.

Env

VariableDefaultPurpose
VIBIE_API_BASEhttps://vibie.ioOverride for local dev (http://localhost:3000)

License

MIT

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.