How do I convert an ODT file to PDF in Node.js without LibreOffice?

Use odf-kit's built-in Typst emitter: import odtToTypst from 'odf-kit/typst', pass the .odt bytes, and write the returned .typ string to disk. Then compile to PDF with the Typst CLI: typst compile document.typ document.pdf. odf-kit is zero-dependency and works in any JavaScript environment — no LibreOffice installation required.

Is there an npm package to convert ODT to PDF?

Yes. odf-kit includes a built-in ODT-to-Typst emitter: npm install odf-kit, then import odtToTypst from 'odf-kit/typst'. It converts any .odt file to Typst markup which can be compiled to PDF with the Typst CLI. odf-kit is the only JavaScript library with a built-in ODT-to-Typst emitter.

What is Typst and why use it for PDF generation?

Typst is a modern open-source document compilation system that produces high-quality PDFs from plain text markup. Unlike LibreOffice, it starts in milliseconds, runs as a single binary with no installation wizard, works on all platforms, and produces clean, well-typeset output. The Typst CLI can be installed via npm, Homebrew, winget, or system package managers.

Can I convert ODT to PDF in a serverless environment?

Yes. The odf-kit Typst emitter (odtToTypst) runs anywhere JavaScript runs — Node.js, browsers, Deno, Bun, Cloudflare Workers. Compiling the .typ output to PDF requires the Typst CLI, which is a single small binary that works on Linux (including Lambda and container environments). The Typst WASM build also enables fully in-browser PDF compilation.

Convert ODT to PDF in JavaScript — No LibreOffice Required

The pipeline

odf-kit reads the .odt file and emits Typst markup. The Typst CLI compiles it to PDF. odf-kit handles the hard part — Typst handles the rendering.

.odt file

→

odtToTypst()

→

.typ markup

→

typst compile

→

.pdf

Try it online — no CLI needed

Convert your ODT file to Typst markup in your browser, then compile to PDF at typst.app. Free, private — your file never leaves your device.

Open the ODT to PDF tool →

Quick start

Install odf-kit and the Typst CLI:

# Install odf-kit
npm install odf-kit

# Install the Typst CLI (choose one)
npm install -g typst          # via npm
brew install typst            # macOS
winget install --id Typst.Typst  # Windows

Convert a .odt file to PDF:

import { odtToTypst } from "odf-kit/typst";
import { readFileSync, writeFileSync } from "node:fs";
import { execSync } from "node:child_process";

const bytes = new Uint8Array(readFileSync("document.odt"));

// Step 1 — convert .odt to Typst markup
const typ = odtToTypst(bytes);
writeFileSync("document.typ", typ);

// Step 2 — compile to PDF
execSync("typst compile document.typ document.pdf");

// document.pdf is ready

Parse once, emit to multiple formats

If you already have a parsed OdtDocumentModel — for example from readOdt() — use modelToTypst() directly to avoid re-parsing the file. This is the preferred approach when generating both HTML and PDF from the same document:

import { readOdt } from "odf-kit/reader";
import { modelToTypst } from "odf-kit/typst";
import { readFileSync, writeFileSync } from "node:fs";
import { execSync } from "node:child_process";

const bytes = new Uint8Array(readFileSync("document.odt"));

// Parse once
const model = readOdt(bytes);

// Emit to HTML
const html = model.toHtml({ fragment: true });
writeFileSync("document.html", html);

// Emit to Typst, compile to PDF
const typ = modelToTypst(model);
writeFileSync("document.typ", typ);
execSync("typst compile document.typ document.pdf");

What gets converted

ODT element	Typst output	Status
Headings (levels 1–6)	`= H1`, `== H2`, …	✓
Paragraphs	Block text with blank-line separation	✓
Text alignment	`#align(center)[…]`	✓
Bold	`text`	✓
Italic	`_text_`	✓
Underline	`#underline[text]`	✓
Strikethrough	`#strike[text]`	✓
Superscript / subscript	`#super[]` / `#sub[]`	✓
Hyperlinks	`#link("url")[text]`	✓
Font color	`#text(fill: rgb("…"))[…]`	✓
Font size	`#text(size: 14pt)[…]`	✓
Font family	`#text(font: "…")[…]`	✓
Highlight color	`#highlight(fill: rgb("…"))[…]`	✓
Unordered lists	`- item`	✓
Ordered lists	`+ item`	✓
Nested lists	Two-space indent per level	✓
Tables	`#table(columns: N, …)`	✓
Table column widths	`columns: (5cm, 10cm)`	✓
Footnotes / endnotes	`#footnote[…]`	✓
Bookmarks	`<label>`	✓
Page number field	`#counter(page).display()`	✓
Page count field	`#counter(page).final().first()`	✓
Named sections	Comment header + body	✓
Tracked changes	`#underline[]` / `#strike[]` / transparent	✓
Page geometry	`#set page(width:, height:, margin: (…))`	✓
Images	Comment placeholder (see below)	Placeholder

Images in the PDF

Typst does not support inline base64 image data without filesystem access. Each image in the source .odt is emitted as a comment placeholder at its document position:

/* [image: logo 10cm × 6cm] */

To include images in the final PDF, extract the image data from the document model and write the files alongside the .typ output before compiling:

import { readOdt } from "odf-kit/reader";
import { modelToTypst } from "odf-kit/typst";
import { readFileSync, writeFileSync } from "node:fs";
import { execSync } from "node:child_process";

const bytes = new Uint8Array(readFileSync("document.odt"));
const model = readOdt(bytes);

// Extract images from the model and write them to disk
let typ = modelToTypst(model);
for (const node of model.body) {
  if (node.kind === "paragraph") {
    for (const span of node.spans) {
      if ("kind" in span && span.kind === "image" && span.name) {
        const ext = (span.mediaType ?? "image/png").split("/")[1];
        const filename = `${span.name}.${ext}`;
        writeFileSync(filename, Buffer.from(span.data, "base64"));
        // Substitute the placeholder with a real #image() call
        typ = typ.replace(
          `/* [image: ${span.name}`,
          `#image("${filename}") /* was:`
        );
      }
    }
  }
}

writeFileSync("document.typ", typ);
execSync("typst compile document.typ document.pdf");

Why not shell out to LibreOffice?

LibreOffice headless (libreoffice --headless --convert-to pdf) is the traditional approach. It works, but comes with real costs in production:

Startup time — LibreOffice takes 2–5 seconds per conversion even for small documents
Deployment size — LibreOffice is several hundred megabytes and dominates Docker image size
Serverless incompatibility — LibreOffice doesn't run on AWS Lambda, Vercel, Cloudflare Workers, or similar environments without significant workarounds
Security surface — spawning a process with filesystem access from your application is a meaningful attack surface
Fragility — headless LibreOffice is notoriously unreliable in containerized environments; fonts, locale, and display settings all affect output

The Typst CLI is a single small binary that starts in milliseconds, produces deterministic output, and runs cleanly in containers. odf-kit's Typst emitter adds zero new JavaScript dependencies.

Comparison with other approaches

Approach	No LibreOffice	No subprocess	Serverless	npm install
odf-kit + Typst CLI	✓	Typst only	✓ (container)	✓
LibreOffice headless	✗	✗	✗	✗ OS install
docxtemplater + LibreOffice	✗	✗	✗	Partial
Pandoc (ODT → PDF)	✗ Needs LaTeX	✗	✗	✗ OS install

Typst CLI installation options

# via npm (works on all platforms)
npm install -g typst

# macOS (Homebrew)
brew install typst

# Windows (winget)
winget install --id Typst.Typst

# Linux (many distros — check repology.org/project/typst)
sudo apt install typst       # Debian/Ubuntu (if available)
sudo pacman -S typst          # Arch Linux

# Verify installation
typst --version

ESM setup

odf-kit is ESM-only. Your script must use import syntax. Either add "type": "module" to your package.json, or use a .mjs file extension:

// package.json
{
  "type": "module"
}

// convert.mjs — or convert.js with "type": "module" above
import { odtToTypst } from "odf-kit/typst";
import { readFileSync, writeFileSync } from "node:fs";
import { execSync } from "node:child_process";

const [,, input, output] = process.argv;
const bytes = new Uint8Array(readFileSync(input));
writeFileSync("out.typ", odtToTypst(bytes));
execSync(`typst compile out.typ ${output}`);

odf-kit does more than convert

The same library that converts .odt files to PDF can also create .odt files from scratch, fill existing templates with data, and read .odt files as structured HTML. One install, one dependency.

Convert .odt to HTML in JavaScript — extract content for web display
Fill .odt templates with JavaScript — design in LibreOffice, fill from code
Generate .odt files in Node.js — build documents from scratch
Generate .odt without LibreOffice — lightweight alternative overview

Convert ODT to PDF in JavaScript
without LibreOffice

The pipeline

Quick start

Parse once, emit to multiple formats

What gets converted

Images in the PDF

Why not shell out to LibreOffice?

Comparison with other approaches

Typst CLI installation options

ESM setup

odf-kit does more than convert

Convert your first .odt to PDF

Convert ODT to PDF in JavaScriptwithout LibreOffice

The pipeline

Quick start

Parse once, emit to multiple formats

What gets converted

Images in the PDF

Why not shell out to LibreOffice?

Comparison with other approaches

Typst CLI installation options

ESM setup

odf-kit does more than convert

Convert your first .odt to PDF

Convert ODT to PDF in JavaScript
without LibreOffice