Think inside your AI world.

Tinybird through Unl

Tinybird's MCP server will hand back the exact p95 on any endpoint, pulled straight from its own service data. Whether that number is fast enough to put in front of a customer is a threshold you've already set once, the hard way.

Tinybird through Unl reads live endpoint-latency data against the launch-SLA rule you have ratified, so 'is this endpoint fast enough to launch?' comes back conditional with the condition it hasn't met.

What Tinybird holds

Query Tinybird's MCP server about a Workspace and it surfaces:

  • SQL results run against the Tinybird Query API
  • the list of Data Sources in a Workspace, including service Data Sources
  • published API Endpoints with their names, parameters, and descriptions
  • natural-language-to-SQL translation via text_to_sql
  • service data on endpoint calls and performance, such as pipe_stats_rt

What the naked read gives you

Tinybird's MCP server runs the query against the Query API, lists every published endpoint with its name and parameters, lists a Workspace's data sources including the service data sources that track usage and performance, and can turn a plain question into SQL through text_to_sql. It reports exactly what an endpoint is doing right now, without saying whether that performance clears it for a launch.

The frame judges the data it is given; it does not verify the source’s accuracy.

What changes when Tinybird is read measured

A founding engineer at a real-time quoting tool ratified a launch rule in Unl: no customer-facing endpoint ships to the pricing page until its p95 response time sits under three hundred milliseconds over the trailing day, because a slow quote endpoint once caused checkout abandonment during a launch nobody caught in time.

'Is this endpoint fast enough to launch?' returns, measured: not yet - the service data source read shows the get_quote endpoint's p95 sitting at four hundred and ninety milliseconds over the last twenty-four hours, above the three-hundred-millisecond ceiling. Tinybird supplied the latency figures; your ratified rule supplied the ceiling.

And back again

The fix is on the endpoint side - trim the underlying pipe's SQL or add a materialised view, and Tinybird's own service data shows the drop in p95 next time anyone asks. Unl reads that improvement into the same page, not a fresh one.

The answer comes back measured against what you already decided, and why.

The lane is live and open to this tool today: one box, paste anything. If it speaks MCP, Unl can reach it. Readings arrive unprompted, the data beside the criterion; Unl is a courier, not a warehouse, and keeps only your keys and the frame.

Questions people ask

What would move this endpoint to launch-ready?

Getting p95 latency under the three-hundred-millisecond ceiling, typically by optimising the underlying pipe's SQL or adding a materialised view. The same ratified rule reads the improved figure and clears the endpoint.

How do I connect Tinybird to Claude?

Enable Tinybird's remote MCP server for your Workspace and add it to Claude's MCP configuration with a Tinybird auth token scoped to the resources you want visible - Unl reads through whatever that token permits.

Does Unl write to my workspace?

No. Unl calls Tinybird's execute_query, list_datasources, and list_endpoints tools to see current state; publishing or changing an endpoint is something you do yourself.

What this is

Think inside your AI world — you stay in command

Unlimitless (Unl to friends) holds what you've settled, reads what your tools are showing, and catches what's changed out in the world — and hands your AI whatever bears on the work, the moment it's needed, without you asking. The right thing, in front of the model, unprompted, with you in command of the call. So you keep moving toward what you set out to build, on top of everything you've already decided.

It plugs into Claude, Claude Code, ChatGPT and Cursor as an MCP connector. Quick to connect, in a couple of steps.

Unlimitless is open now to invited Alpha. Apply for the Beta waitlist to come in ahead of the full launch:

Alpha is invite-only · free at launch.