MCP Integration

MCP lets your favorite AI helper (like Claude or Cursor) talk to Kayse AI for you. You ask a question or give a task, and the AI reads or updates your data — all in plain English. 🤖

What is MCP?

MCP stands for Model Context Protocol. Think of it as a shared language that AI tools and apps can use to talk to each other.

With MCP turned on, you can say things like:

• "Show me all open cases."
• "Add a note to case 123 that the client wants a callback."
• "Send a quick email to the client on case 456."

The AI does the work inside Kayse AI for you. No clicking around. No copy and paste. ✨

Which AI Tools Work?

Any AI tool that supports MCP will work. Popular ones include:

• Claude Desktop
• Cursor
• Cline

If your AI tool says it "supports MCP," you can connect it to Kayse AI.

What Can the AI Do?

Once MCP is set up, your AI helper can:

• Look things up — cases, clients, case lists, forms, campaigns, calls, tasks, notes, admins, and audit logs.
• Add or update records — create clients, update a case, write a note, or change a case status order.
• Do common actions — send a text or email tied to a case, run a webhook test, or replay a past webhook delivery.

Safety First

MCP cannot delete things for you. It can only read, add, or update. This keeps your data safe. 🔒

Set It Up (Step by Step)

Step 1 — Get Your API Key

An API key is like a password that lets the AI tool log in for you.

1. Open Company Settings → API Keys in Kayse AI.
2. Create a new key and copy it.
3. Keep it private — treat it like a password. Don't share it or post it online.

Step 2 — Add Kayse AI to Your AI Tool

Claude Desktop

1. Open the Claude Desktop settings file:
   • Mac: ~/Library/Application Support/Claude/claude_desktop_config.json
   • Windows: %APPDATA%\Claude\claude_desktop_config.json
2. Add this block (paste your own key where it says your_api_key_here):

{
  "mcpServers": {
    "kayse": {
      "url": "https://public-api.kayse.ai/v1/mcp",
      "headers": {
        "Authorization": "Bearer your_api_key_here"
      }
    }
  }
}

3. Save the file and restart Claude Desktop.

Cursor

1. Open your Cursor MCP settings (.cursor/mcp.json in your project, or the global Cursor settings).
2. Add the same block:

{
  "mcpServers": {
    "kayse": {
      "url": "https://public-api.kayse.ai/v1/mcp",
      "headers": {
        "Authorization": "Bearer your_api_key_here"
      }
    }
  }
}

3. Save and restart Cursor.

Step 3 — Try It

Ask your AI helper a simple question, like:

• "Show me my most recent 5 cases."
• "List my active campaigns."

If it gives you Kayse AI data, you're all set! 🎉

Prompts to Try First

• "Show me recent open cases."
• "Create a note on case 123 that says the client asked for a callback tomorrow."
• "List the webhook logs for webhook 42."
• "Send an email to the client on case 123 with a follow-up message."
• "Show me the current case statuses in the right order."
• "List admins and recent audit-log activity for client 456."

Who Can the AI See?

The AI uses your API key. That means it sees the same things your key can see — no more, no less.

If your key can't open a case or webhook, the AI can't either. Your company admins set this up in Company Settings.

Troubleshooting

The AI tool says it can't connect.

• Check that your API key is still active in Company Settings.
• Make sure the URL and Authorization line in the config are spelled right.
• Restart your AI tool after you change the config.

The AI says it doesn't have permission.

• Your API key may be missing a permission. Ask your admin to update it.

Learn More

• MCP API Reference — Full details for developers.
• API Authentication — How to make and manage API keys.
• Model Context Protocol — The official MCP site.