Tinybird Data Platform MCP. Audit your real-time data infrastructure via chat.

Tinybird Data Platform MCP Server connects your agent directly to a real-time analytics engine in Tinybird. You're getting admin-level access, letting you manage everything from raw data ingestion right through complex transformations and ad-hoc queries—all without ever having to open the dashboard yourself.

This isn't just another wrapper; it lets you audit your whole analytical infrastructure using nothing but natural conversation. You can inspect every layer of data processing, check usage metrics on a whim, and run full workflows instantly. It’s like having an expert teammate who knows where all the switches are.

Listing and Inspecting Your Data Sources

You need to know what data you're working with? You can start by running list_workspaces to see every workspace the admin token has access to. From there, use list_datasources to pull a complete list of every single source in your current workspace. If you wanna dig deeper into one specific source, the agent uses get_datasource_details to fetch all the schema and metadata for that data set.

It’s how you verify what columns exist and what type of data they hold.

To check usage stats on a data source—things like row counts or total storage size—you just call get_datasource_stats. You get immediate metrics showing exactly how much data's sitting there. You can also list all the authentication tokens defined in the workspace using list_auth_tokens, which lets you audit who has access to your critical data.

Mapping Out Your Data Pipelines (Pipes)

The core of any analytics stack is its pipes, and this tool gives you full visibility. You can call list_pipes to get a list of every defined transformation pipeline in the workspace. If you need to understand how one specific pipe works—its history, configuration, or logic flow—you use get_pipe_details. This reveals the inner workings of that data process.

Want to see exactly which steps make up a complex pipe? The agent uses list_pipe_nodes to list all the individual SQL nodes that compose any given transformation pipe. It lets you trace the flow from start to finish, making debugging way faster than clicking through menus. You can run predefined data workflows by executing a full Pipe definition using query_pipe_data, and it returns the resulting dataset as structured JSON, ready for your agent to use.

Running Queries and Auditing Security

Need some immediate answers? The agent doesn't rely on pre-built pipes; you can run arbitrary SQL queries directly against your Tinybird data workspace using execute_sql_query. You just type the query, and it gets instant results. If your task is to audit your security setup, listing all authentication tokens via list_auth_tokens gives you a clear rundown of every key available.

Basically, you can list all workspaces (list_workspaces), then list all data sources (list_datasources) and get the schema details for any one source (get_datasource_details). You'll also check usage metrics like row counts or storage size using get_datasource_stats. For transformations, you can list all defined pipes (list_pipes), get a deep dive into a pipe’s configuration via get_pipe_details, and even see every individual node in the process with list_pipe_nodes.

When you're done inspecting, you can run a whole pipeline using query_pipe_data or fire off an immediate, custom query with execute_sql_query. It’s everything your data stack needs, right here.