Renderers

captchakit ships three renderers out of the box. They all implement the same Renderer protocol (content_type + async def render), so you can swap them without touching the manager.

Renderer	Output	Deps	Use when
ImageRenderer	PNG bytes	Pillow	default, universal
SVGRenderer	SVG bytes	stdlib	HD screens, smaller payload
AudioRenderer	WAV bytes	stdlib	accessibility alternative

SVGRenderer

from captchakit import SVGRenderer, Theme

renderer = SVGRenderer(
    width=220,
    height=80,
    font_size=40,
    theme=Theme.DARK,
    font_family="sans-serif",
)

• Pure Python string templating — no Pillow dependency.
• Typical payload is ~1 kB (vs ~4–8 kB for the equivalent PNG).
• Scales losslessly on high-DPI screens.
• Reuses the same Theme palette / noise settings as ImageRenderer.

Serve as Content-Type: image/svg+xml. Browsers render it inline inside an <img> tag just like a PNG.

Picking the right renderer

• Default → ImageRenderer. Max compatibility; every email client, image widget and ancient browser can display it.
• Mobile-first or HD webapps → SVGRenderer. Smaller over-the-wire, crisp at any zoom level.
• Accessibility → always pair one of the visual renderers with AudioRenderer (see Audio).

Bring your own

Implement the Renderer protocol — a content type plus an async render(challenge) -> bytes:

class QRRenderer:
    content_type = "image/png"
    async def render(self, challenge):
        import qrcode, io
        buf = io.BytesIO()
        qrcode.make(challenge.prompt).save(buf, format="PNG")
        return buf.getvalue()