Python + Rust · MIT licensed

Your data. Your design. Your PDF.

Turn Python data and HTML/CSS into invoices, reports, and print documents. Start with one install, then grow from a single PDF to a compiled variable-data workflow.

Make your first PDF Explore examples

python -m pip install fullbleed

Python 3.10–3.14 · Windows, macOS, Linux

A real generated PDF. Open it · Get the source

Self-contained wheels   /   Explicit fonts and assets   /   Repeatable output   /   Free for commercial use under MIT

A small first step

Save this as hello.py, run python hello.py, and open invoice.pdf.

from pathlib import Path
import fullbleed

html = "<h1>Invoice INV-1042</h1><p>Consulting: USD 1,200.00</p>"
css = "@page { size: A4; margin: 20mm; } h1 { color: #175c52; }"
pdf = fullbleed.PdfEngine().render_pdf(html, css)
Path("invoice.pdf").write_bytes(pdf)

Pick a document to build

• Invoices from your data

  Turn JSON or CSV into itemized invoices. Keep data, layout, and font assets explicit.

  Invoice guide →

• Reports that flow across pages

  Use headings, tables, page margins, headers, and footers for a document that grows with its content.

  See a five-page report →

• One template, many records

  Bind stable fields or let changing content reflow through a compiled template.

  Variable-data guide →

• Tagged and print-oriented output

  Author semantic content, inspect the output, and retain evidence for the profile checks you run.

  Accessibility → · Print output →

Built for a document pipeline

The wheel bundles a Rust rendering engine, a Python API, a CLI, and fonts. It requires no third-party Python runtime packages. You can render, preview, inspect, and verify in the same workflow.

Fullbleed uses static HTML/CSS as its layout language. Read the CSS coverage for your templates, and the tool selection guide when you also need live browser rendering or general PDF editing.

Open source, with inspectable evidence

Fullbleed is MIT licensed. The 2.4.0 release includes downloadable wheels and retained engineering evidence. The performance report describes specific measured workloads and their limits.

Read the Python API · Set up a coding agent · Report an issue