Forum Discussion

Divyanshu15's avatar
Divyanshu15
Copper Contributor
Oct 27, 2026

Seeking Guidance: Microsoft Copilot Integration with Apache Superset via MCP

Seeking Guidance: Microsoft Copilot Integration with Apache Superset via MCP

Hi everyone,

We are using Apache Superset 6.1.0 with PostgreSQL and ClickHouse datasets and exploring integration with Microsoft Copilot through Superset's MCP server.

Our objective is to allow users to ask questions in natural language and receive text-based answers, tabular results, and AI-generated charts or dashboards based on our existing data.

I would appreciate guidance on the following:

  1. Integration feasibility: Can Microsoft Copilot Studio connect to Superset's MCP server and support this use case?
  2. Capabilities and performance: Can it handle natural-language queries, data retrieval, chart generation, and dashboard creation effectively?
  3. Licensing: Which Microsoft plan is suitable for this use case? Would a trial or pay-as-you-go option be sufficient for a proof of concept?
  4. Costs: What are the estimated licensing, message-consumption, and Azure infrastructure costs for a small POC and production deployment?
  5. Security: Can we enforce authenticated access and read-only permissions while protecting our PostgreSQL and ClickHouse data?

Our initial goal is to validate the integration through a small proof of concept before committing to a paid plan.

If anyone has implemented a similar integration, I would appreciate your recommendations, practical experience, and estimated costs.

Thank you!

copilot studio
microsoft copilot

1 Reply

  • Jamony's avatar
    Jamony
    MCT
    Nov 05, 2026

    You can prototype natural-language access to your Superset data in Copilot Studio, provided its MCP endpoint supports the required Streamable HTTP transport. MCP support alone does not establish Superset version compatibility or guarantee its chart and dashboard capabilities.

     

    Enable generative orchestration, then select Tools > Add a tool > New tool > Model Context Protocol. Configure the endpoint and supported authentication, then test the exposed tools against known queries. Measure answer accuracy and latency; tables, charts, and dashboard creation depend on what those tools actually return or perform.

     

    Use authenticated connections and enforce read-only access in the server and underlying databases; agent instructions alone should not enforce permissions.

     

    A trial supports testing in the test panel, but cannot publish. Pay-as-you-go can fund a published POC. Estimate Copilot Credits from measured usage and budget MCP hosting separately; reliable production costs require expected traffic and infrastructure details.