MCPifex Quickstart: Connect an MCP Server in 5 Steps

This MCPifex quickstart gets an AI app connected to a real MCP server in about five minutes. You pick a server, choose what the AI may do with it, test it, get an API key, and paste one line into Claude Code, ChatGPT or Cursor. There's nothing to install. MCPifex runs the server for you.

What you'll have at the end

An AI assistant that can use a real tool. If you connect PostgreSQL, for example, you can ask Claude "which tables do I have, and how many rows are in orders?" and it answers from your actual database. Write access stays off unless you turn it on.

Before you start

• A free MCPifex account. Register here; no card needed.
• An MCP client, such as Claude Code, Claude Desktop, ChatGPT or Cursor.
• The login details for whatever you're connecting, like a database host, username and password.

Browse the marketplace to see which servers you can connect.

Step 1: create an instance

In the portal, pick a server and create an instance of it. An instance is your own configured copy, with its own credentials, tool settings and API key. You can make several of the same server, for example one PostgreSQL instance per database.

Step 2: choose which tools the AI may use

Each server lists its tools. Read-only tools start switched on. Anything that writes or deletes, like a SQL execute, starts switched off.

The gateway enforces this on every call. Your AI app only ever sees the tools you enabled, and a call to any other tool is refused.

Step 3: test the connection

Click Test connection. A wrong password or missing permission shows up now, not halfway through your first conversation.

Step 4: generate an API key

Once the test passes, generate an API key. It starts with mcpx_. Keep it private.

Step 5: connect your AI app

Every app connects to the same address, https://mcp.mcpifex.com/mcp. Only the way it sends your key differs.

Claude Code

Run this once in your terminal:

#!/bin/sh
# MCPifex — Claude Code CLI snippet.
# Replace <YOUR_MCPX_KEY> with your MCPifex API key (starts "mcpx_"), from
# the portal's API key page, then run this once from any shell that has the
# `claude` CLI on PATH.
claude mcp add mcpifex --transport http https://mcp.mcpifex.com/mcp \
  --header "Authorization: Bearer <YOUR_MCPX_KEY>"

Then check it with claude mcp list. You should see mcpifex marked as connected.

ChatGPT and other apps that can't send headers

Some apps have nowhere to paste an Authorization header. For those, put the key in the URL instead. Replace <YOUR_MCPX_KEY> with your key:

1. On ChatGPT web, open Plugins, select the plus button, and choose Add custom MCP server. If you do not see it, turn on developer mode first under Settings → Security and login → Developer mode. Menu names change.
2. Enter the name MCPifex and a short description.
3. Set Connection to the public URL, including /mcp: https://mcp.mcpifex.com/mcp/<YOUR_MCPX_KEY>.
4. Set authentication to None: the MCPifex key in the URL authenticates the gateway request.
5. Accept the risk warning and select Create, then review the discovered tools.
6. In a conversation, type @ in the prompt box and select the app, then test one named read-only tool.

Older accounts may show Settings → Connectors → Advanced → Add connector instead.

These steps follow the OpenAI connect-ChatGPT documentation. OpenAI's help article lists Business, Enterprise and Edu workspaces on the web; some guides also report Plus and Pro. Check Settings in your own account.

Treat this URL like a password: anyone who has it can use every tool enabled on that key. Use the header form wherever your app supports it.

Test from the command line (optional)

Want to check an instance works before involving an AI app? This script connects with curl and lists the tools your key can use:

#!/bin/sh
# MCPifex — raw MCP handshake over curl (initialize + tools/list).
# Replace <YOUR_MCPX_KEY> with your MCPifex API key (starts "mcpx_"), from
# the portal's API key page. Requires curl only.
set -e

GATEWAY_URL="https://mcp.mcpifex.com/mcp"
MCPX_KEY="<YOUR_MCPX_KEY>"

# 1. initialize — the gateway replies with an Mcp-Session-Id response header
#    that every later request on this session must echo back.
INIT_HEADERS=$(mktemp)
INIT_BODY=$(mktemp)
curl -sS -D "$INIT_HEADERS" -o "$INIT_BODY" \
  -X POST "$GATEWAY_URL" \
  -H "Authorization: Bearer $MCPX_KEY" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "initialize",
    "params": {
      "protocolVersion": "2025-06-18",
      "capabilities": {},
      "clientInfo": { "name": "mcpifex-curl-example", "version": "1.0.0" }
    }
  }'

SESSION_ID=$(grep -i '^mcp-session-id:' "$INIT_HEADERS" | tr -d '\r' | cut -d' ' -f2-)
echo "Mcp-Session-Id: $SESSION_ID"
cat "$INIT_BODY"
echo
rm -f "$INIT_HEADERS" "$INIT_BODY"

# 2. tools/list — filtered to the tools enabled for this API key.
curl -sS -X POST "$GATEWAY_URL" \
  -H "Authorization: Bearer $MCPX_KEY" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -H "Mcp-Session-Id: $SESSION_ID" \
  -d '{
    "jsonrpc": "2.0",
    "id": 2,
    "method": "tools/list",
    "params": {}
  }'
echo

After you connect

• Your AI app sees exactly the tools you enabled in step 2, and nothing else.
• Each session gets its own isolated server, which stops when the session ends or goes idle.
• Every call is logged, so you can see in the portal what the AI actually did.

The free plan covers 3 instances. Paid plans go up to 30, and tool calls are unlimited on every plan. See pricing for details.

← Back to docs