#Connect your AI to Clickmax

For your AI to genuinely operate Clickmax, it needs two pieces that go together:

• MCP (Model Context Protocol) — gives the AI real-time access to your account's tools and data (leads, transactions, funnels, members…).
• Skills — teach the AI to use Clickmax: the rules, concepts and flows of each area.

#Prerequisites

• a client that supports remote MCP over HTTP (Claude Code, Cursor, Codex, etc.);
• a valid Bearer token for your Clickmax account (step 1);
• a terminal with npx available (step 3).

#Step 1 — Create your MCP token

Generate a token in your account:

The token is prefixed with cx-token-:

cx-token-d803b54c48f98fad159f13db8349ddd6

#Step 2 — Connect MCP

The server uses Streamable HTTP:

https://mcp.clickmax.io/mcp

Every request needs the authentication header:

Authorization: Bearer <SEU_TOKEN_MCP>

Pick the tab for your client and swap <SEU_TOKEN_MCP> for the token from step 1:

claude mcp add --transport http clickmax "https://mcp.clickmax.io/mcp" \
  --header "Authorization: Bearer <SEU_TOKEN_MCP>"

/mcp

{
  "mcpServers": {
    "clickmax": {
      "url": "https://mcp.clickmax.io/mcp",
      "headers": {
        "Authorization": "Bearer <SEU_TOKEN_MCP>"
      }
    }
  }
}

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "clickmax": {
      "type": "remote",
      "enabled": true,
      "url": "https://mcp.clickmax.io/mcp",
      "headers": {
        "Authorization": "Bearer <SEU_TOKEN_MCP>"
      }
    }
  }
}

[mcp_servers.clickmax]
url = "https://mcp.clickmax.io/mcp"
bearer_token_env_var = "CLICKMAX_MCP_TOKEN"

export CLICKMAX_MCP_TOKEN="cx-token-..."

{
  "mcpServers": {
    "clickmax": {
      "url": "https://mcp.clickmax.io/mcp",
      "headers": {
        "Authorization": "Bearer <SEU_TOKEN_MCP>"
      }
    }
  }
}

#Step 3 — Install the Skills

The public Agent Skills teach the AI to operate each area of the platform. Install them all with one command:

npx skills add https://docs.clickmax.io

The command discovers and installs the skills from the docs discovery index (/.well-known/agent-skills/index.json).

The base URL on its own, to paste into the command:

https://docs.clickmax.io

Prefer to install manually? Download the package with every skill already organised into folders (<skill>/SKILL.md + reference files):

#Step 4 — Test

With MCP connected and the skills installed, ask for a real task:

Am I connected to Clickmax? Which account am I?

Identify the hot leads with no contact in the last 7 days.

If everything is right, the AI returns your identity, reads real data from the account and calls the skill that matches the task. If the client does not load the skill on its own, ask for it by name (e.g. Use the clickmax-pipelines skill…).

#Skills catalogue

There are 17 public skills, grouped by area. The phrase "Use when" is the trigger the AI uses to load the skill — describe your task in those terms to call it up.

#CRM

#Payments and offers

#Members

#Automations and funnels

#For agents — programmatic discovery

Endpoints served by the docs for clients compatible with Agent Skills:

Each index entry carries name, description, type (skill-md or archive), url and digest (sha256) for integrity checking.

#Troubleshooting

#Security good practice

• Never commit or share the token in repositories, screenshots or logs.
• Prefer environment variables when the client supports interpolation in the header.
• Keep the server named clickmax to make it easier to identify and support.
• Revoke the token immediately if you suspect it has been exposed, and generate a new one.