Control Seat V0.1.2 — Smarter AI & Query Tags
V0.1.2 is a big one. The AI editor is substantially faster and more reliable, especially on busier pages. A new Query Tags source lets you back any tag — or a whole data table — with a live SQL, Flux, or PromQL query. And the flows editor picked up a round of UX polish so working on a flow feels the same as working on a page.
What's in V0.1.2
A faster, smarter AI editor
The AI assistant gets things right more often, on the first try, and finishes noticeably faster. Multi-step requests like "add two charts, style them like the existing one, and wire them to pump RPM" stay on task instead of drifting after the first step.
- Faster on typical edits — most simple changes now land in one round trip.
- Fewer misses — the assistant double-checks its own work against what you asked for, and self-corrects when it misses. If you asked for two charts and it only added one, it notices and fixes it instead of stopping.
- Better sense of spacing and layout — the assistant can now measure the gaps between elements on the page, so requests like "make it match the spacing of the row above" actually match.
- Binding-aware — it can see and edit existing data bindings on a component (not just the HTML), so "stop this chart from updating live" or "unbind this from pump temp" work the way you'd expect.
- Stop button — cancel a running AI edit mid-flight from the popover or edit modal, instead of waiting for it to finish.
- Live status — the popover tells you what the assistant is doing right now ("measuring spacing", "adding chart 2 of 2", "verifying result") instead of a loading dot.
Query Tags — tags and tables backed by live queries
Any tag can now be backed by a query against a datasource. Point it at a Postgres / MySQL / SQL Server table, an InfluxDB / Flux query, or a PromQL expression, and the tag's value updates as the query returns new results.
- Query Tag source in the tag form — pick a datasource, write the query in the built-in editor, preview the result, save.
- A proper query editor with syntax highlighting for SQL, Flux, and PromQL. SQL autocompletes table and column names from the connected datasource, so you can discover your schema as you type instead of alt-tabbing to another tool.
- Table preview right in the tag modal — for queries that return multiple rows or columns, see the full result before you save.
- Data tables bind to query tags — drop a data table on the canvas, pick a query tag, and the table renders its live results. Only query tags show up as options for data tables, so you can't accidentally bind a table to a scalar tag.
- Consistent query editor everywhere — the same editor is now used across tag sources, chart data queries, and data tables, with matching theming and keyboard shortcuts.
Flows editor now feels like the pages editor
The flows editor picked up the same save / publish / version-history behavior the pages editor has had, so there's less to relearn when switching between them.
- Save auto-publishes live flows — when a flow is running, clicking Save pushes your changes live, the same way saving a published page does. No more second click in the Publish popover to ship an update.
- Publish popover simplified — a stopped flow shows Publish (first publish + enable in one click); a running flow shows Unpublish. The old "Publish changes" / "Republish" label was confusing and has been removed.
- Delete old versions — flow version history now supports deleting old snapshots, matching the pages editor. Publishing a flow also no longer auto-creates a version row every time; history is snapshot-on-demand, so your version list doesn't fill up with noise.
JSON tag cleanup
JSON-valued tags (tags whose value is an object or array) now work consistently everywhere — alarms, historian queries, chart tag bindings, custom-chart snapshots, and the tag panel all handle JSON tag paths the same way. If you had a JSON tag that worked in one place but not another, that's fixed.
More AI and flows work landing soon. Read the Flows docs or get in touch for a demo.