Przejdź do treści
Qamera AI
Wypróbuj za darmo

Setup

Create an API key with the right permissions and connect Cursor, Claude Desktop or another MCP client to Qamera AI.

What you need

  1. A Qamera AI account — sign up at qamera.ai.
  2. An API key with two permissions, created in your workspace settings.

There is nothing to install. The MCP server is hosted by us at https://mcp.qamera.ai; your AI client connects to it over the network.

Creating your API key

  1. Sign in to Qamera AI.
  2. Open your workspace Settings and go to API Keys.
  3. Choose Create key and give it a name you will recognise later, such as the tool you are connecting.
  4. Select both permissions:
    • Register Content (content.register) — everything to do with products, generation, previews and approvals.
    • Read Content Status (content.read) — checking whether a generation finished.
  5. Copy the key when it is shown.

Nothing is pre-selected, and one permission is not enough. The form requires at least one and offers no default, so it is easy to create a key with only the first. Such a key works for almost everything and then fails on status checks with a permissions error — which looks like a generation that never started rather than a key that is missing a box.

What the key looks like. It begins with mk_live_, then an identifier, a dot, and a secret: mk_live_<identifier>.<secret>. If what you pasted has no dot in it, or does not start with mk_live_, the copy was incomplete.

The secret is shown once. It cannot be recovered afterwards — if you lose it, create a new key and delete the old one. Keep it out of shared documents and out of anything you commit to a repository.

Cursor

Create or edit .cursor/mcp.json in your project:

{
  "mcpServers": {
    "qamera-ai": {
      "url": "https://mcp.qamera.ai",
      "transport": { "type": "sse" },
      "headers": {
        "Authorization": "Bearer YOUR_QAMERA_AI_API_KEY"
      }
    }
  }
}

Replace YOUR_QAMERA_AI_API_KEY with the key you copied. Keep the word Bearer and the single space that separates it from the key.

Then:

  1. Save the file.
  2. Restart Cursor so it picks up the new server.
  3. Open the AI chat and ask something like "List my products in Qamera AI."

If the tools do not appear, see Troubleshooting — the most common cause is the key rather than the configuration.

Claude Desktop

Claude Desktop connects to a remote server as a connector, added in its settings rather than in a configuration file.

  1. Open Settings — Ctrl + , on Windows, Cmd + , on macOS — then Connectors.
  2. Choose Add, then Add custom connector.
  3. Enter the server URL: https://mcp.qamera.ai
  4. Open Request headers and add:
    • header name x-api-key
    • value: your key, exactly as you copied it
  5. Click Add, then enable the connector in a conversation through the + button.

One caveat, and it may affect you. Header authentication in custom connectors is a beta feature that Anthropic is rolling out gradually. If you do not see a Request headers section in the dialog, your account does not have it yet — Claude Desktop cannot authenticate to Qamera AI without it, and Cursor or another client is the route for now. Anthropic's own connector documentation is the current source on availability.

On Team and Enterprise plans an owner adds the connector once under Organization settings, and members then connect to it individually.

Other MCP clients

Any MCP-compatible client can connect to the hosted server.

  • Server address: https://mcp.qamera.ai
  • Authentication: send the key in one of these headers:
    • Authorization: Bearer YOUR_API_KEY
    • X-Api-Key: YOUR_API_KEY

Checking the server is up

curl https://mcp.qamera.ai/health

A successful answer means the server is reachable. It says nothing about your key — a rejected key fails at connection time, which the troubleshooting page covers.

If you configured Qamera AI before August 2026

The earlier address — an IP with a port number, over plain http — still works, so nothing you have set up will stop working today. Move to https://mcp.qamera.ai when convenient: the old address sends your API key unencrypted, and it will be withdrawn once integrations have moved.

Next steps