Skip to content

Automatic translation — the Portuguese version is the reference. View in Portuguese

Updated on 2026-10-03

Editor, pages, layouts and themes

How the dashboard editor works inside — a canvas with snapping and fine adjustment, multi-select, undo, pages, TV mode, draft × published, layout per device and themes.

On this page (8)

The editor is 100% web (no desktop app). You build the dashboard on a grid canvas, with drag and drop, undo/redo, autosave, versions and themes. Examples use the synthetic workspace Comércio Aurora Ltda.

Looking for a button?

This page explains the concepts. The name of each button is in The editor ribbons, The visual pane, Fields, measures and variables and The right-click menu.


Canvas: drag, drop and resize#

  1. What it is — a grid screen (48 columns) where you position and resize visuals.
  2. When to use it — every time you build/edit a dashboard.
  3. How it works — in Tools → Move and resize, you drag the card by its middle and resize it by the corner and side handles. The card lands exactly where you dropped it: the position goes to the nearest grid cell (48 columns) plus a fine adjustment, without jumping. Snapping shows a guide when the card's edge or center lines up with another card, one axis at a time; hold Alt to turn snapping off. A click without dragging doesn't move anything. Dragging a field onto the canvas creates a visual; dragging a field into a visual puts it in the right place (axis/value/legend). Finish layout saves and exits the mode.
  4. Prerequisites — a dashboard (it starts empty, as a draft) and a model with fields.
  5. How to validate — drag produtos.categoria onto the canvas (it creates a visual), then vendas.valor into it; move the card until the snapping guide appears, drop it and confirm it stayed where you left it.

Move and resize mode with the snapping guide aligning two cards
6. Limits and caveats — automated tests cover the grid and card-drop logic; the final appearance depends on the browser. 7. Sample data (synthetic) — a "Revenue by category" visual built from produtos.categoria + vendas.valor. 8. Evidence — the grid, drop and anti-overlap logic (pure functions) is covered by a deterministic suite; DOM rendering has no visual test.


Alignment, multi-select and copy/paste#

  1. What it is — align/distribute, Z order, lock, hide; multi-select (⌘-click); copy/paste across pages and dashboards. The commands are in the right-click menu (Position, Size, Arrange) and in the Selection pane on the Tools ribbon.
  2. When to use it — arranging several visuals at once, reusing a block.
  3. How it works — pure (tested) layout operations; the clipboard uses localStorage (Ctrl+C / Ctrl+V), including across dashboards.
  4. Prerequisites — two or more visuals.
  5. How to validate — select three visuals with ⌘-click, align left, copy and paste them on another page; confirm they arrive together.
  6. Limits and caveats — the clipboard depends on the browser (localStorage).
  7. Sample data (synthetic) — a "Sales KPIs" block (3 KPIs) copied from the "Summary" page to "Detail".
  8. Evidence — align/distribute and copy-paste operations (pure layout functions) covered by a deterministic suite.

Undo/redo and autosave#

  1. What it is — an undo/redo stack (20 steps) and automatic saving with debounce.
  2. When to use it — always; quick recovery from an editing mistake.
  3. How it works — 100% optimistic undo/redo (Ctrl+Z / Ctrl+Shift+Z); autosave writes only if the layout changed.
  4. Prerequisites — a dashboard in edit mode.
  5. How to validate — move a visual, press Ctrl+Z and confirm it goes back; wait for the debounce and reload the page to see the persisted state.
  6. Limits and caveats — the stack belongs to the editing session (20 steps). Switching pages, inserting or removing a visual and the page's Arrange aren't part of Undo (removing always asks for confirmation).
  7. Sample data (synthetic) — move "Revenue by category" and undo.
  8. Evidence — undo/redo stack and automatic saving (writes only if the layout changed) covered by a deterministic suite.

Pages, presentation and focus#

  1. What it is — multiple pages/tabs (with a ceiling), full-screen/presentation mode (←/→ switches pages), FOCUS mode (1 visual) and TV mode (pages that rotate on the big screen).
  2. When to use it — separating "Summary" and "Detail"; presenting on a big screen.
  3. How it works — page tabs with a ceiling; tooltip pages are hidden from the tab bar (used as tooltip pages — see Filters and interactions).
  4. Prerequisites — a dashboard with at least one page.
  5. How to validate — create the "Summary" and "Detail" pages, enter presentation mode and navigate with the arrow keys.
  6. Limits and caveats — tooltip pages stay out of the tab bar. For a big screen that rotates pages on its own, use TV mode (Page → Layout → TV ▾ and the TV link under Share) — see Layout per device and TV mode.
  7. Sample data (synthetic) — "Summary" (KPIs) and "Detail" (vendas table) pages.
  8. Evidence — pages, presentation mode and focus mode covered by the editor's tested logic.

Draft × published and versioning#

  1. What it is — separation between the draft (what you edit) and the published version (what readers see), with version control and restore.
  2. When to use it — editing without affecting consumers; going back to a previous version.
  3. How it works — readers see the frozen version; there's a pending-changes diff between draft and published; up to 20 versions (512KB each) with a per-widget diff, restore and automatic pruning.
  4. Prerequisites — edit/publish permission.
  5. How to validate — publish, edit the draft, see the pending-changes diff; restore a previous version and confirm it comes back.
  6. Limits and caveats — a ceiling of 20 versions/512KB; better than a binary .pbix for diffs, but it isn't a Git integration.
  7. Sample data (synthetic) — v1 "3 KPIs" → v2 "3 KPIs + table"; restore v1.
  8. Evidence — publishing (draft×published), pending-changes diff and version history covered by a deterministic suite; history on the /versoes page.

Layout per device: desktop, tablet, phone and TV#

  1. What it is — a dedicated arrangement for each device: Desktop (from 1024px), Tablet (640 to 1023px), Phone (below 640px) and TV (TV link, 16:9 screen).
  2. When to use it — dashboards viewed on a phone, on the store's tablet or on a big screen.
  3. How it works — in Page → Layout, choose the device and arrange the cards; the desktop layout doesn't change. Without a dedicated arrangement, the phone stacks the visuals in desktop order, the tablet draws like the desktop and the TV uses the desktop layout scaled. Starting points: Auto stack (phone) and Copy from desktop (tablet and TV), both with confirmation.
  4. Prerequisites — an existing dashboard, in edit mode.
  5. How to validate — arrange the phone, open the public link on a phone (or in a narrow window) and check the arrangement; go back to Desktop and check that nothing changed there.
  6. Limits and caveats — on touch, the regular tooltip opens, but the tooltip page only appears with a mouse. It's responsive web; there's no native app/push (that's Roadmap).
  7. Sample data (synthetic) — "Summary" with 3 KPIs stacked on the phone and side by side on the tablet.
  8. Evidence — device choice by width, auto-stacking, copy from desktop and protection of the phone layout covered by a deterministic suite.

Full walkthrough in Layout per device and TV mode.


Themes and templates#

  1. What it is — the dashboard theme (a palette of 6 validated hex colors, font, background, card style) and templates via JSON import/export.
  2. When to use it — standardizing the visual identity; reusing a dashboard's structure without carrying data.
  3. How it works — colors are sanitized against injection; there's a contrast check and ready-made Brazilian templates; the template never carries data (structure only), with anti-__proto__ protection on import.
  4. Prerequisites — nothing beyond the dashboard.
  5. How to validate — change the palette and confirm the change on every visual; export the template, import it into a new dashboard and confirm it starts with no data.
  6. Limits and caveats — there's no template gallery at creation time: the templates are JSON files, under Page → Theme and template ▾ (Download theme, Import theme, Save template, Use template, New from template).
  7. Sample data (synthetic) — the "Aurora" theme (6 colors) exported and reimported.
  8. Evidence — color sanitization against injection, contrast check, Brazilian templates and protected import (anti-__proto__) covered by a deterministic suite.

Empty / loading / error states#

  1. What it is — "No data" cards, an error card with the reason, loading skeletons and a sample badge when the result was truncated.
  2. When to use it — automatic; ensures that one visual's failure doesn't take down the dashboard.
  3. How it works — a query error becomes a card with the clean reason (cleanQueryError strips the datalake's technical noise); truncation propagates the "sample" mark.
  4. Prerequisites — none.
  5. How to validate — force a visual with no data and confirm "No data"; force an error and confirm the card with a reason instead of a broken screen.
  6. Limits and caveats — the error text is sanitized for reading; it isn't the technical stack trace.
  7. Sample data (synthetic) — a visual on vendas filtered to a period with no rows → "No data".
  8. Evidence — error card with a clean reason, loading skeletons and propagation of the "sample" mark covered by a deterministic suite.

Last reviewed: 2026-10-03.

Related links