Connecting an AI Agent

https://*.claudemcpcontent.com

Connect from the Claude Connector Directory

Sisense is listed in Claude's connector directory. This is the quickest way to connect on Managed Cloud.

1. In Claude, open Customize > Connectors, and click + next to Connectors.

2. Find Sisense and click Connect.

3. Enter your Sisense deployment URL. The directory listing uses the base endpoint (https://mcp.sisense.com/mcp, with no target), so Sisense first shows a screen where you enter your deployment URL. See Option 2 under Managed Cloud above.

4. Sign in with your Sisense credentials and approve access.

Add a Custom Connector

Add Sisense as a custom connector instead in these cases:

• Self-hosted deployments. The directory listing connects to the Sisense-hosted endpoint. Use your own instance's endpoint (https://<your-deployment>/mcp).

• Administrators who want to set the deployment up front. Use the endpoint with the target parameter (Option 1 above), so users connect straight to your deployment without the deployment URL screen.

On individual plans:

1. Open Customize > Connectors > + Add > Add custom connector.

2. Give it a name (for example, Sisense) and paste your endpoint URL.

3. Save. Claude prompts you to sign in the first time you use it.

https://*.web-sandbox.oaiusercontent.com

Connect from the ChatGPT Directory

Sisense is listed in ChatGPT's directory. This is the quickest way to connect on Managed Cloud.

1. Open the Sisense listing in ChatGPT, or find Sisense in ChatGPT's directory.

2. Click Connect.

3. Sign in with your Sisense credentials and approve access. If Sisense asks for your deployment URL, enter it first.

Add a Custom Connector

On self-hosted deployments, add Sisense as a custom connector with your own instance's endpoint (https://<your-deployment>/mcp). The directory listing connects to the Sisense-hosted endpoint.

1. Open Settings > Connectors.

2. Add a custom connector, name it (for example, Sisense), and paste your endpoint URL.

3. Save, then sign in to Sisense when ChatGPT prompts you.

Custom connector availability depends on your ChatGPT plan.

Add the server to your MCP configuration. On Managed Cloud:

{
  "mcpServers": {
    "sisense": {
      "url": "https://mcp.sisense.com/mcp?target=<your-deployment-url>"
    }
  }
}

On self-hosted deployments:

{
  "mcpServers": {
    "sisense": {
      "url": "https://<your-deployment>/mcp"
    }
  }
}

Reload the MCP servers. Cursor opens a browser sign-in the first time the server is used.

Any client that supports remote MCP servers over HTTP with OAuth can connect. Add the endpoint as a remote/HTTP server (not a local command). The client discovers Sisense's OAuth configuration automatically and registers itself. No manual client ID or secret is required.

Reload or restart the MCP client, and confirm it supports remote MCP servers over HTTP. Check that the endpoint is entered as a remote/HTTP server, not a local command.

Ask your administrator to check the following:

• The sisenseMcp setting is in the Sisense Helm override values (for example, installer/extra_values/helm/sisense.yml), not in config.yaml or cluster_config.yaml. It must be nested under global:, with enabled: true, and the installer or upgrade must be run afterward. See Enabling the MCP Server on Self-Hosted Deployments.

• The oauth-gateway and mcp-service pods are running.

• The Enabled toggle is on under System Configuration > MCP.

• If multitenancy is enabled on the deployment and the endpoint returns the Sisense "Page not Found" page, this is a known limitation that is fixed in Sisense 2026.4.1.

See Enabling the MCP Server on Self-Hosted Deployments.

Connections stay signed in for 3 days from sign-in, and using the connection does not extend this. After 3 days, you must sign in to Sisense again. Depending on the client, it either prompts you to sign in, or tool calls fail until you reconnect Sisense in the client's connector settings. Sign in again, then retry the request.

If the problem persists after you sign in, confirm you are signing in to the correct Sisense instance.

Data source names are case-sensitive. Ask the agent to list your data sources and use the exact name.

Check whether your client supports MCP Apps. Clients without it receive the narrative insights rather than a rendered chart or dashboard, and nothing is wrong with the connection or the query. In Claude or ChatGPT, if the chart or dashboard area stays blank, or you get text only, see the CORS entries under Interactive Charts and Dashboards in Claude and ChatGPT above.

Your administrator must allow the client's content domain in Sisense CORS settings. See Interactive Charts and Dashboards in Claude and ChatGPT above. Other tools still work without it, but answers are text only.

Charts and dashboards shown in a conversation stay viewable for about 24 hours. Ask the agent to rebuild it to bring it back.

Be more specific: name the chart type, the fields, and any sorting or filter. Simplify the request, then build it back up.

The AI-powered tools need Cloud-Linked Features and a configured LLM provider. Ask your administrator to confirm Cloud-Linked Features is enabled under Admin > Sisense Intelligence > Feature Management, and that an LLM provider is set up: either the LLM managed by Sisense, which draws on your Sisense credit package, or your own provider under Provider Configuration. If you use the Sisense-managed LLM, also check that your Sisense credit package has not run out and that credits are allocated to your tenant and user group under Admin > Sisense Intelligence > Credits & Allocations. The discovery tools (getDataSources and getDataSourceFields) do not need this and keep working.