Download the PHP package paperdoc-dev/paperdoc-lib without Composer

Paperdoc Library

A zero-dependency PHP library for generating, parsing and converting documents — PDF, HTML, CSV, DOCX, XLSX, PPTX, Markdown and more.

Features

• Generate documents from scratch (PDF, HTML, CSV, DOCX, XLSX, PPTX, Markdown)
• Parse existing documents into a unified in-memory model
• Convert between any supported formats in one call
• Rich document model — typed headings, ordered/bullet lists (nested), bookmarks, code blocks, blockquotes, images, tables, page breaks and typed document properties (author, subject, dates…)
• Per-page layout (v0.7.0+) — per-section PageSetup with custom size (or any PageSize enum), padding, full-page background image (cover / contain / auto / stretch since v0.7.1) or color ; absolutely-positioned TextZone blocks with clip / ellipsis / visible overflow strategies and per-paragraph alignment (left / center / right / justify, v0.7.1) ; document-wide running headers / footers with {page} / {pages} / {title} / {date} / {datetime} placeholders
• Typographic quality (v0.7.3) — per-glyph metrics from the 14 standard PDF fonts (Core 14, AFM-derived) so centering, right-alignment, justification and word-wrapping are pixel-accurate ; ascent-aware vertical stacking (no more eyebrow/title overlap) ; justification combines Tw (word-spacing) and Tc (character-spacing) and falls back to flush-left when a line would produce visible "rivers"
• Layout & typography APIs (v0.8.0) — per-section header/footer override (Section::setHeader/setFooter/hideHeader/hideFooter) ; Section::setVerticalAlignment(TOP/CENTER/BOTTOM) for chapter openers / colophons ; per-side padding shortcuts (setPagePaddingTop/Right/Bottom/Left) ; ParagraphStyle::setFirstLineIndent() (CSS text-indent) ; TextStyle::setLetterSpacing() (PDF Tc operator, HTML letter-spacing) ; new HorizontalRule block element rendered to PDF (stroked line), HTML (<hr>), Markdown (---) and DOCX (bordered paragraph)
• Native rendering core — every block element renders cleanly to DOCX, PDF, HTML and Markdown: typed headings (<h1>/<w:pStyle>), nested lists (<ul>/<w:numPr>), blockquotes, code blocks (with language hint), bookmarks, embedded or on-disk images
• Hyperlinks — parse <w:hyperlink> from DOCX and round-trip them to HTML <a>, Markdown [text](url) and DOCX hyperlink relationships, with anchors and tooltips
• Batch processing — open and process multiple files at once
• Laravel integration — first-class ServiceProvider and Facade
• AI-powered features via Neuron AI (OCR, LLM extraction)
• Typed exceptions — ParserException, RendererException, UnsupportedFormatException, InvalidDocumentException all extending a common PaperdocException
• Zero native binary dependencies — pure PHP

Requirements

Optional (Laravel)

Installation

Laravel auto-discovery

The PaperdocServiceProvider and Paperdoc facade are registered automatically via Laravel's package auto-discovery.

Quick Start

DocumentManager uses static methods (create, save, open, …). A document is built from Section instances: use addSection($section) or addSection() to append an empty section, or openSection() when you want a fluent chain (addParagraph, addHeading, …) on the new section. Bold and other run styles live on TextStyle. For advanced layouts (custom page size, full-page backgrounds, absolutely-positioned text zones, running headers/footers), see Page layout below.

Standalone PHP

Laravel (via Facade)

Supported Formats

Document Model

Every format shares the same strongly-typed in-memory structure:

All block elements implement Paperdoc\Contracts\BlockElementInterface. Styles live in Document/Style/ (ParagraphStyle, TextStyle, TableStyle), links in Document/Link/TextLink, typed document properties in Document/Metadata.

Example — build a richly-typed document

Page layout, text zones, headers & footers

Available since v0.7.0. Each section can declare its own page geometry through a PageSetup value object, and place absolutely-positioned TextZone blocks anywhere on the page. Combine several sections (each with its own PageSetup) to build documents where every page has a different size and background. Add a global RunningElement to the Document to draw a header/footer on every page.

Configure a page

Section exposes shortcut setters (setPageSize(), setPageDimensions(), setPagePadding(), setPageBackgroundImage(), setPageBackgroundColor()) that delegate to a lazily-created PageSetup.

Fit the background image — cover / contain / auto / stretch

Available since v0.7.1. Both renderers (PDF and HTML) honour the same four CSS-like modes. cover and auto automatically clip the overflow with a clip path in the PDF and overflow: hidden in the HTML output.

Any other CSS-valid string ('50% auto', '300pt 200pt', …) is accepted as-is in HTML output.

Place text precisely with TextZone

Coordinates use the top-left convention (x=0, y=0 is the top-left of the page) for both PDF and HTML — the PdfRenderer flips to PDF's bottom-left origin internally.

Per-paragraph alignment inside a zone — v0.7.1

Each paragraph of a TextZone carries its own ParagraphStyle, so you can mix several alignments in the same zone (centred title, justified body, right-aligned signature, …):

In the PDF, justification is implemented with the native PDF word-spacing operator (Tw); the last line of a paragraph is intentionally left-aligned to avoid stretched short lines.

Document-wide headers and footers

Supported placeholders in the template: {page} (1-indexed current page), {pages} (total pages), {title} (the document title), {date} (Y-m-d) and {datetime} (Y-m-d H:i). The renderer resolves them per page so you don't need to update the template between pages.

The HTML renderer adds a translucent rgba(255, 255, 255, 0.85) backdrop with a backdrop-filter: blur(2px) behind the running elements so they remain legible on top of any background image. The library does not automatically reserve vertical space for the header/footer — keep that in mind when positioning a TextZone close to a page edge.

Per-section header / footer override (v0.8.0)

A document-level header/footer applies uniformly to every page by default. Often that's not what you want — a cover page should NOT carry the page-number footer (it would either disappear under the artwork or fight with the imagery for legibility), and a colophon on the last page might want a different label. Since v0.8.0, sections can override or suppress the document-level running elements:

Resolution rule — for every page, the renderer picks (in order) :

1. If Section::hideHeader() / hideFooter() was called → no header/footer is drawn.
2. If Section::setHeader() / setFooter() was called with a non-null element → that element is drawn.
3. Otherwise → the document-level header/footer is drawn (if any).

Both PDF and HTML renderers honour this resolution.

Vertical alignment of section content (v0.8.0)

By default the content of a section flows from the top padding downwards. For pages that should breathe vertically — chapter openers, colophons, frontispieces, single-paragraph "blank" pages — you can centre or bottom-anchor the content :

Implementation notes :

• The PDF renderer captures the section's content slice, measures its rendered height, then wraps it in a native PDF q ... 1 0 0 1 0 dy cm ... Q translation block — no overhead, no pre-render measurement pass.
• The HTML renderer applies the same semantics via flexbox (display:flex; justify-content:center for CENTER, flex-end for BOTTOM).
• Sections that overflow onto a second page automatically fall back to TOP alignment to avoid a stale CTM bleeding across pages. If you need centring on a section with lots of content, reduce its content first.

Per-side padding shortcuts (v0.8.0)

Section::setPagePadding(...$values) already accepts CSS-shorthand 1-/2-/3-/4-value forms. When only ONE side needs tweaking — typical for a frontispiece title that should sit ~110pt from the top — the new per-side shortcuts are clearer :

Available : setPagePaddingTop(), setPagePaddingRight(), setPagePaddingBottom(), setPagePaddingLeft(). They all return $this for chaining.

First-line indent and letter-spacing (v0.8.0)

Two style-level additions that previously had to be hacked at the application layer :

Paragraph first-line indent

ParagraphStyle::setFirstLineIndent(float $points) mirrors the CSS text-indent property : only the first line of the paragraph starts further to the right. Negative values produce a hanging indent (first line jutting OUT to the left of the block — a typographic device used in lists or dictionary entries).

The PDF wrap engine sees a tighter budget for the first line so wrapping accounts for the indent correctly. The HTML renderer emits text-indent: Xpt on the paragraph's inline style.

Run letter-spacing

TextStyle::setLetterSpacing(float $points) opens out a run of text by the requested number of points between every pair of adjacent glyphs. Common uses : an opened-out section heading (setLetterSpacing(1.5)), or marking an acronym in small-caps style. Negative values pull glyphs closer.

PDF emits the native Tc operator and resets to 0 after the run, so copy-paste from the rendered file gives back the original (un-spaced) text — a property the previous "insert thin spaces between every glyph" workarounds did not have. HTML emits letter-spacing: Xpt. measureTextWidth() and wrapText() correctly account for letter-spacing.

Horizontal rule (v0.8.0)

A first-class block element for visual separators :

Renderers :

• PDF : a stroked horizontal line, drawn with the requested thickness and stroke colour.
• HTML : <hr> with inline CSS (border-top:Wpt solid C ; width:X ; margin:T 0 B).
• Markdown : --- thematic break (CommonMark).
• DOCX : the canonical Word "horizontal line" — an empty paragraph carrying a <w:pBdr><w:bottom .../></w:pBdr> with the requested colour and thickness in eighths-of-a-point.

Rendering

Since v0.5.0, every element of the document model is natively rendered by all four core renderers — no element is silently dropped, every output is a valid file format.

php artisan vendor:publish --tag=paperdoc-config bash composer test

or

./vendor/bin/phpunit

src/ ├── Concerns/ # Shared traits ├── Console/ # Artisan commands ├── Contracts/ # DocumentInterface, ParserInterface, BlockElementInterface… ├── Document/ # Core model (Document, Section, Paragraph, Heading, ListBlock, Bookmark, CodeBlock, Blockquote, Metadata…) ├── Enum/ # Format enums ├── Exceptions/ # PaperdocException + typed exceptions ├── Facades/ # Laravel Facade ├── Factory/ # Document/Parser factories ├── Llm/ # AI/LLM integration (Neuron AI) ├── Ocr/ # OCR integration ├── Parsers/ # Format-specific parsers ├── Renderers/ # Format-specific renderers ├── Support/ # DocumentManager and helpers └── PaperdocServiceProvider.php

---

## Contributing

We welcome contributions! Please read [CONTRIBUTING.md](CONTRIBUTING.md) before opening a pull request.

### Contributors

Thanks to everyone who has contributed to **paperdoc-lib**. A full list is kept in [CONTRIBUTORS.md](CONTRIBUTORS.md).

- **Olivier Mourlevat** — [@olivM](https://github.com/olivM) — DOCX hyperlink parsing, HTML/Markdown hyperlink rendering ([#4](https://github.com/paperdoc-dev/paperdoc-lib/pull/4))

---

## Changelog

See [CHANGELOG.md](CHANGELOG.md) for release history.

---

## License

Paperdoc Library is released under the **[MIT License](LICENSE)** — free to use, modify and distribute, commercial or not.

© Paperdoc — [paperdoc.dev](https://paperdoc.dev)