Fonts & images

Fonts

Text embeds as real glyph runs with subset fonts. It stays selectable, searchable, and copyable. Pass a URL or raw bytes. You can also use the googleFonts helper:

import { render } from "takumi-pdf";

import { googleFonts } from "@takumi-rs/helpers";

const pdf = await render(doc, {
  fonts: [...(await googleFonts(["Inter"])), { name: "Brand Sans", weight: 700, data: fontBytes }],
  fontFamilies: ["Brand Sans", "Inter", "sans-serif"],
});

fontFamilies sets the fallback chain for each render. Registered fonts deduplicate across calls that use the same renderer.

There is no system font to fall back on. A character that none of the registered fonts covers would draw nothing and leave nothing in the text layer, so render rejects it and names the character and its codepoint. Register a font covering every script the document uses. Watch the Google Fonts subsets in particular: latin and latin-ext split their coverage, so one file on its own can leave common accented characters out.

Images

The renderer never fetches over the network. Pass pre-fetched bytes for each image URL in the document:

import { render } from "takumi-pdf";

const logoBytes = new Uint8Array(await (await fetch(logoUrl)).arrayBuffer());

const pdf = await render(doc, {
  images: [{ src: logoUrl, data: logoBytes }],
});

SVG image sources embed as vector paths and gradients, so logos stay sharp at any zoom. Subtrees that vectors cannot express (filters, embedded bitmaps) rasterize at 2× their placed size. Raster sources embed once for each distinct pixel data.

Runtimes

takumi-pdf is wasm-only. It runs on Node.js, Bun, and Cloudflare Workers. The module fits within the Workers free-tier size limit. Import takumi-pdf/no-init to control wasm instantiation.