Is any plan to launch DS mcp?
bump
Hi Thanks for the question!
Do you need the MCP servers for the current Data Studio API that supports only basic asset management, or do you need AI to finish more tasks on Data Studio?
Could you share a few examples of your use cases?
Hi!! I just want to manage/create dashboards
By āmanageā, do you mean manage the sharing settings, e.g.
- Add/remove viewers of a existing report or datasource.
By ācreateā, do you mean create it from your external data, e.g.
- Create a datasource for the data you have (e.g. in Google Sheets)
- Then create a report on top of the datasource.
If convenient, could you list a few concrete use cases in your mind. Ideally, also add the related business value for you.
Regardless, the team is actively working to enhance the integration of Data Studio with AI agents. Any input or feedback you can provide would be highly valued and greatly appreciated.
Hii!! Iām looking to use MCP for more than just data extraction. I want to fully manage the dashboard itself things like access permissions, data sources, tweaking dimensions and metrics, and editing the charts and pages. Most BI tools out there only use MCP to query data or create measures, so having a tool that actually handles the āmanagementā side would be a great way to differentiate from the rest
I think it would be a cool to integrate conversational analytics into dashboards. If the agent could automatically pull the context from the dashboard like the specific queries, tables, running on the backend or connect under one chatbot the multiple pages a dashboad can have
Hi hd,
One example use case I have in mind: helping to manage the data source definitions (adding/updating metrics) for example.
Right now, a coding agent helped me to create a dbt project from the ground up. I have my marts and dimensions, and would like to create a few metrics inside the GDS. So far, I am using the UI to create my metrics.
Thank you for the feedback!
They can be roughly categorized to
- Report/datasource permission management.
- Datasource creation.
- Chart creation.
- Report based analysis.
In your use cases, do you want to do them inside Data Studio, e.g. through a Gemini sidebar, or do you want to let other external agents (e.g. Gemini, Claude Cowork etc) to operate Data Studio? Which one is more important for you and why?
Hi Jice_Lavocat, welcome to the community and thank you for your feedback!
Why do you think agent based datasource creation and management is important? Is it one of the non-trivial or time-consuming part of you project? Would you like share a concrete example?
Do you need this feature inside Data Studio, or through external agents (e.g. Gemini, Claude Cowork, OpenClaw etc)?
Hi HD, I agree with your categorization
For my use cases, I need the MCP to function externally. Iām looking to integrate it into my current projects and pipelines, so limiting it to a sidebar or specific UI would restrict its utility (IMHO)
The reason is scalability. By exposing these capabilities via an MCP server, I can integrate them into my existing workflows
Noted! Thank you!
Hi! It would be very very helpful if an MCP / API dumped the outputs so that LLMs can read the plots.
Do you mean the final data used by a chart, e.g. a 2d table view of the used dimensions and metrics?
Could you provide a concrete use case on this, how is the data in your datasource prepared, what datasource is added to Data Studio? How do you plan to use this feature with AI agents?
I mean the graphs plotted on a chart in Data Studio, whatever the datasource is or however it is prepared.
It would be great to automate LLM reviews of the data e.g. monthly reviews of the dashboard: an LLM would request the plots in the dashboard and be able to make a summary, alerting of meaningful changes.
Noted, and fully agree. Thanks!