Connect Qlik MCP in Gemini CLI

[nico19505]

Has anyone already connected the Qlik MCP to Gemini CLI?

I’m trying to retrieve Qlik information within my Gemini setup.

I created a user with the appropriate permissions, as described in the Qlik Help documentation, to connect Gemini to Qlik using OAuth.

I then generated a token using the client ID and client secret of the user, and added it to the Gemini configuration like this:

{
  "mcpServers": {
    "qlik-cloud": {
      "httpUrl": "https://<MY_TENANT_QLIK>.qlikcloud.com/api/ai/mcp",
      "headers": {
        "Authorization": "Bearer <MY_TOKEN_OAUTH>"
      }
    }
  }
}

However, when I run /mcp list, Gemini indicates that Qlik Sense is disconnected.

Has anyone experienced this issue before or managed to get it working?

[meri45clark]

Most likely issue is auth/token or MCP handshake, not your JSON.

Quick checks:

Make sure the OAuth token has correct scopes + audience for /api/ai/mcp

Test endpoint directly:

curl -H "Authorization: Bearer <TOKEN>" https://<TENANT>.qlikcloud.com/api/ai/mcp
If this returns 401/403 → Qlik config/token issue
If it works → Gemini CLI config issue (header/format)
Try removing headers in MCP config to isolate auth behavior
Also ensure token is not expired (very common) ezcardinfo com

If /mcp list shows “disconnected”, it usually means handshake failed, not just missing data.

[marksouzacosta]

Hi @nico19505 ,

I don't think Qlik MCP Supports Bearer Authentication and I don't think Gemini Supports Qlik MCP Server.

Regards,

Mark Costa

[Leigh_Kennedy]

I don't know on the Google side, 

But here is my config that works with LM Studio:

{
  "mcpServers": {
    "Remote_Qlik_MCP": {
      "url": "https://abc.ap.qlikcloud.com/api/ai/mcp",
      "auth": {
        "CLIENT_ID": "myid",
        "CLIENT_SECRET": "my_secret"
      }
    }
  }
}

In the OAUTH config, I have:

Client type: web
Redirect URLS:  http://127.0.0.1:33389/mcp-oauth-callback 

and the rest is as recommended in the docs.

[agigliotti]

Hi @nico19505 ,

I suggest you to use API Key to authenticate Gemini CLI vs Qlik MCP.

Best Regards

[Daniele_Purrone]

@marksouzacosta 

The web based version of Gemini does not support MCP servers.
However, the CLI version does.

[marksouzacosta]

@Daniele_Purrone, I was referring to the oAuth authentication. Qlik MCP only works oAuth authentication and Gemini CLI does not - at least as far as I have tested.

[Daniele_Purrone]

@marksouzacosta , I currently don't have a Gemini CLI setup to test it myself, but I've heard internally that it is possible to connect.
And here's the Gemini page about Oauth Support: https://geminicli.com/docs/tools/mcp-server/#oauth-support-for-remote-mcp-servers 

[Laurent_Cornilleau]

I confirmed that it works. A an alternative you can use Antigravity 2.0 (IDE or CLI). Even if Google will still continue to support Gemini CLI for Enterprise customers (N.B it will be stopped for for Google AI Pro and Ultra on June 18th), they recommand to move to Antigravity CLI.

[marksouzacosta · Accepted solution]

Thank you @Daniele_Purrone and @Laurent_Cornilleau. It is finally working on my side now.

On Qlik Cloud side
-------------------------------------------------

Create an oAuth Client and add the following callback URLs:

https://antigravity.google/oauth-callback
http://localhost:7777/oauth/callback

For reference, I have used these properties in my oAuth Client:
Client Type: Native
Scopes: user_defaul and mcp:execute
Published: no
Consent method: required

On Windows side
-------------------------------------------------

Antigravity

Config file location:
C:\Users\<you>\.gemini\config\mcp_config.json

MCP Server:

{
  "mcpServers": {
    "data-voyagers-qlik-mcp": {
      "serverUrl": "https://<your-tenant>.us.qlikcloud.com/api/ai/mcp",
      "oauth": {
        "clientId": "<your-client-id>"
      }
    }
  }
}

Result:

Antigravity IDE

Shares Antigravity configuration. Only need to re-authenticate.

Gemini CLI

Config file location:
C:\Users\<you>\.gemini\settings.json

MCP Server:

{
  "mcpServers": {
    "data-voyagers-qlik-mcp": {
      "url": "https://<your-tenant>.us.qlikcloud.com/api/ai/mcp",
      "oauth": {
        "clientId": "<your-client-id>",
        "redirectUri": "http://localhost:7777/oauth/callback"
      }
    }
  }
}

Open your Gemini CLI and type:

/mcp auth <name-of-your-mcp-server>

In my example:

/mcp auth data-voyagers-qlik-mcp

Result:

Regards,

Mark Costa