Skip to main content
A poll floats on a slide as a card, or fills the screen as its own step. Both take the same poll; the difference is what the room looks at.

​
Where a card goes

The rect is fractions of the slide: x and y for the top-left corner, w and h for the size. Left out, the card lands in the bottom-right quarter, the house default.
Three rules, from watching what looks right on a big screen:
  1. Cover a screenshot, a picture or empty space. Never text.
  2. If the default corner holds text, take another corner, or the picture’s area.
  3. A card under its look’s minimum (30% of the slide’s width and 45% of its height for most looks, more for a few) is unreadable from the back of the room, and gets a warning.
The API refuses a card that runs off the slide. Look at the result with get_screenshot before handing the deck over. Its warnings name a card that is too small to read, cut off, or on top of another card or text box, and coversPageText lists the words of a PDF page that a card hides. Move a card with update_overlay, sending only its new rect. With annotate, get_screenshot draws every element’s box and name on the picture, and the ones that clash in red.
card-covers-text

On the slide's text

The card sits on both text boxes, so all three are drawn in red and warnings names each overlap.
card-clear-of-text

Clear of it

The same card in the house corner. Nothing clashes, so every box stays blue.

​
How a card looks

The style is the card itself: the chart variant, the accent, the background, the corner radius, shadow and outline, and what the card prints. Leave it out for the house look; send only the fields you want to change.
card-look-house

The house look

No style at all.
card-look-donut

Donut

The example above.
card-look-columns

Columns

variant set to columns, a rose accent and radius 24, with showPercent on.

​
The looks

Most widgets offer several looks for their result, chosen with a property on the poll: mcStyle for multiple choice, scaleStyle for a rating, hpStyle for 100 points, and so on. The first look is the default. Some looks need a setting: arc and chord on multiple choice draw what people picked together, so they need multiSelect. The reference lists every look per widget and the rule behind it.

​
Full screen instead

A question that deserves the whole room’s attention, a word cloud at the end, a Q&A, a spin of the wheel, goes in as a poll step with add_step. It fills the screen when the presenter walks onto it.
Any widget can float or fill the screen. A card can print percentages and counts only where the widget draws them; the reference says which.