Templates

Layered creatives with dynamic placeholders

A template is a Konva-based scene graph: a stack of layers (text, image, shape) on a fixed canvas. Layers marked as dynamic become placeholders that you fill at render time via modifications.

Anatomy

Field	Description
id	Stable identifier (tpl_<uuid>)
name	Human-readable name
width, height	Canvas dimensions in pixels
current_version	Latest published version number
is_published	Whether the template can be used by renders
layers	Ordered list of layers (z-index = array order)

Layer types

• Text — string content, font, weight, color, alignment, fit-to-box
• Image — bitmap layer with crop, fit, and tint
• Shape — rectangle, ellipse, line with stroke and fill

Dynamic layers

Mark any layer as dynamic to expose it as a modifications target. Reference by name:

{
  "modifications": [
    { "name": "headline", "text": "Black Friday — 50% off" },
    { "name": "product", "image_url": "https://cdn.example.com/sneaker.png" }
  ]
}

Versioning

Every publish creates a new immutable version. Renders pin to current_version at request time, so old renders remain reproducible.

See the API reference for templates.