The query console

The console is the patch bay. Patch a shape, a metric, a dimension and a set of conditions into one read, and the request that will actually go over the wire is written out beside it before anything is sent.

It is at /app/:site/query. Nothing on it writes to your site.

What you are building

The console composes the same query envelope every screen uses:

• the endpoint: overview, timeseries or breakdown
• the window: a range or explicit start and end, plus a comparison
• the audience: human, agent or all
• the dimension, for a breakdown
• the conditions, as filters with and/or

Two deliberate behaviours

It does not run on its own. A console runs when you tell it to. When the builder has drifted from the last reading, the screen says so rather than quietly re-reading on every keystroke. Cmd/Ctrl + Enter runs it from anywhere on the screen.

The request is shown, not described. The URL on the tape is built by the same serialiser the dashboard uses, so it cannot drift from what is sent. Copy it, paste it into curl with an API key, and you get the same bytes back.

curl "https://analytics.example.com/api/stats/breakdown?site=1&range=30d&audience=agent&dimension=operator" \
  -H "Authorization: Bearer mfg_…"

What it is not

It is not SQL. You cannot write a join, invent a metric or reach a table the API does not expose, and that is on purpose: the audience for this product installs a script tag and does not want to learn a schema. Depth is available; it is never required.

If you need something the envelope cannot express, the honest answer is that it needs to become an endpoint, not a text box.

Getting a result out

Any console result exports as CSV through GET /api/stats/export.csv with the same envelope, and a saved reading can become a panel on a custom dashboard.