mcpField
F
Faxify MCP

Faxify MCP

VerifiedFeaturedRemote
mcp.faxify.comCommunication1Updated 2w ago

About Faxify MCP

Faxify MCP connects AI clients to Faxify’s hosted fax service for sending, receiving, and tracking faxes in the US and Canada. Its local stdio bridge serves clients that cannot connect to remote MCP servers directly, while OAuth clients can connect to the hosted endpoint.

Key features

  1. 01Send and receive faxes
  2. 02Track fax status
  3. 03Hosted HTTP MCP endpoint
  4. 04stdio bridge for local clients
  5. 05JWT or OAuth authentication

Use cases

  1. 01Send a document as a fax from an AI client
  2. 02Check whether a fax was delivered
  3. 03Connect Cursor or Claude Desktop through stdio

MCP support

Exposes Faxify fax capabilities through a hosted MCP endpoint. A stdio bridge forwards requests from local-stdio clients to that endpoint; remote OAuth-capable clients can connect directly.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://mcp.faxify.com/api/v1/mcp
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

Tools

Details

From the project README

@faxify/mcp-client

stdio bridge for the hosted Faxify MCP server. It lets MCP clients that launch local stdio servers (Cursor, Claude Desktop's config file, Continue.dev and others) reach https://mcp.faxify.com/api/v1/mcp over HTTP.

Clients that connect to remote MCP servers over OAuth don't need this package: in ChatGPT (developer mode) or as a Claude custom connector, add https://mcp.faxify.com/api/v1/mcp and sign in (see below).

Resources

Installation

Requires Node.js 18 or later, with npx on your PATH. There is nothing to install: the MCP configurations below run npx -y @faxify/mcp-client, which downloads and starts the bridge.

A global install also works and adds the faxify-mcp-client command:

npm install -g @faxify/mcp-client

Usage

For Cursor

Edit ~/.cursor/mcp.json:

{
  "mcpServers": {
    "Faxify": {
      "command": "npx",
      "args": ["-y", "@faxify/mcp-client"],
      "env": {
        "FAXIFY_MCP_URL": "https://mcp.faxify.com/api/v1/mcp",
        "FAXIFY_JWT_TOKEN": "YOUR_JWT_TOKEN_HERE"
      }
    }
  }
}
For Claude Desktop

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (Mac) or %APPDATA%/Claude/claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "Faxify": {
      "command": "npx",
      "args": ["-y", "@faxify/mcp-client"],
      "env": {
        "FAXIFY_MCP_URL": "https://mcp.faxify.com/api/v1/mcp",
        "FAXIFY_JWT_TOKEN": "YOUR_JWT_TOKEN_HERE"
      }
    }
  }
}

You can also connect Claude without the bridge: see For Claude (custom connector).

For Claude (custom connector)

Claude can connect to the Faxify MCP server directly over OAuth, with no token to copy and no bridge:

  1. In Claude (claude.ai or the desktop app), go to Customize → Connectors, click + and choose Add custom connector
  2. Enter the URL https://mcp.faxify.com/api/v1/mcp
  3. Sign in to Faxify when prompted (Google SSO, Apple SSO, or Email/OTP)

On Team and Enterprise plans, an Owner adds the connector under Organization settings → Connectors, and members then click Connect under Customize → Connectors. Anthropic's current steps: Get started with custom connectors using remote MCP.

For ChatGPT (developer mode)

ChatGPT connects to the Faxify MCP server directly over OAuth, with no bridge. Developer mode is available on the web for Plus, Pro, Business, Enterprise and Education accounts:

  1. On chatgpt.com, open Settings → Security and login and turn on Developer mode
  2. Go to chatgpt.com/plugins, click + and create an app with the MCP URL https://mcp.faxify.com/api/v1/mcp
  3. Sign in to Faxify when prompted:
    • Google SSO: Sign in with your Google account
    • Apple SSO: Sign in with your Apple ID
    • Email/OTP: Enter your email and verify with the one-time password
  4. Once authenticated, the Faxify app is available in your ChatGPT chats

OpenAI's current steps: Developer mode.

For Other MCP-Compatible Applications

Any AI application that supports the Model Context Protocol (MCP) can use Faxify MCP. The configuration follows the same pattern across all MCP-compatible clients.

Common MCP-compatible applications include:

  • Continue.dev (VS Code and JetBrains extension)
  • Other MCP-compatible clients that launch stdio servers

Generic Configuration Pattern:

Most MCP clients use a JSON configuration file with the following structure:

{
  "mcpServers": {
    "Faxify": {
      "command": "npx",
      "args": ["-y", "@faxify/mcp-client"],
      "env": {
        "FAXIFY_MCP_URL": "https://mcp.faxify.com/api/v1/mcp",
        "FAXIFY_JWT_TOKEN": "YOUR_JWT_TOKEN_HERE"
      }
    }
  }
}

Configuration file locations by application:

  • Continue.dev: save the JSON above as .continue/mcpServers/faxify.json in your workspace (or add an mcpServers entry to your Continue config.yaml). The file contains your token, so add .continue/mcpServers/ to your .gitignore, or keep the token in Continue's secrets and reference it as ${{ secrets.FAXIFY_JWT_TOKEN }} in config.yaml.
  • Other tools: Refer to your specific application's MCP documentation

Note: After updating your configuration file, restart the application for changes to take effect.

Environment Variables

  • FAXIFY_MCP_URL - MCP server URL (default: https://mcp.faxify.com/api/v1/mcp)
  • FAXIFY_JWT_TOKEN - JWT token for authentication (get from Settings → MCP Configuration). Required: without it the server rejects the MCP handshake, so no tool is reachable.
  • FAXIFY_MCP_TIMEOUT_MS - How long to wait for the server, in milliseconds (default: 55000). The default answers just before the 60-second tool-call limit in Claude Desktop and other clients built on the MCP SDK, so the model sees why the call stopped. Clients without that limit (such as Cursor) can raise it. If a call times out, it may still complete on the server: check list_faxes or get_fax_status before sending again.
  • FAXIFY_MCP_DEBUG - Set to exactly true to log each tool call's name and arguments to stderr (off by default). The arguments include fax numbers, cover-page text and file links, and your MCP client may keep stderr in its log files, so turn it off when you're done debugging. The JWT is never logged.

Getting Started

First, create an account:

Authentication Methods

For ChatGPT (developer mode) and Claude custom connectors (OAuth, no bridge):

  1. Add the MCP URL https://mcp.faxify.com/api/v1/mcp as a developer-mode app in ChatGPT or as a Claude custom connector
  2. Follow the OAuth authentication flow:
    • Sign in with Google SSO, Apple SSO, or Email/OTP
  3. Once authenticated, start using the Faxify connector directly

For clients that use this bridge (Cursor, Claude Desktop's config file, Continue.dev and other stdio clients):

These clients authenticate with a JWT token:

  1. Log in to Faxify MCP
  2. Go to Settings → MCP Configuration
  3. Copy your JWT token
  4. Add it to your MCP client configuration as FAXIFY_JWT_TOKEN

The token expires, and there is no refresh flow. When tool calls start returning "Authentication Failed", sign in to Settings again, copy the new token into FAXIFY_JWT_TOKEN and restart your client. Settings shows when the current token expires.

How It Works

This wrapper script bridges two different transport mechanisms:

  1. stdio MCP clients (Cursor, Claude Desktop's config file, Continue.dev, etc.) - Use stdio (stdin/stdout) for MCP communication
  2. Faxify MCP Server - Runs as HTTP endpoint on Next.js/Vercel

The wrapper:

  • Reads JSON-RPC requests from stdin (line-delimited JSON)
  • Makes HTTP POST requests to the MCP server
  • Writes responses back to stdout

License

MIT © 2025 Const Agility, LLC

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.