Installation

platex is a two-part system: a compilation service that does the actual LaTeX → PDF work, and a thin client library your app calls. You can also skip the service and compile locally (system TeX Live or the bundled Tectonic binary) — handy for scripts, CLI use, and development.

Install the library

npm install @nandan-varma/platex

The package exposes three entry points — pick the one that matches your runtime:

Import from	Runtime	What it’s for
@nandan-varma/platex	Node.js	compile, createPlatexClient, handleCompileRequest — full library, local-compile fallback included
@nandan-varma/platex/client	Anything with fetch (edge, Workers, Bun, Deno, browsers)	Same client/handler API, remote-only
@nandan-varma/platex/server	Node.js	createApp, createCompileRoute — embed the HTTP API into your own server

See How it works for when to use each.

Set up the service

1. Deploy the platex service. It does the actual compilation and runs anywhere Node.js runs — Vercel, Fly.io, Railway, Render, or your own Docker host. On Vercel it bundles the Tectonic engine into the serverless function.

   git clone https://github.com/nandan-varma/platex
   cd platex
   npx vercel deploy

   Your service is now live at something like https://platex-xxx.vercel.app. See Deploy the service for the full walkthrough, and Self-hosting with Docker for full TeX Live.

2. Point the client at the service. Set PLATEX_SERVICE_URL in your app’s environment (the Vercel dashboard, or .env.local for local dev):

   .env.local

   PLATEX_SERVICE_URL=https://your-platex-service.vercel.app
   # PLATEX_API_KEY=...   # only if you enabled auth on the service

   Both createPlatexClient() and the plain compile() / handleCompileRequest read these automatically — you never thread process.env through your own code.

3. Compile. That’s the whole setup.

   import { compile } from '@nandan-varma/platex';
   
   const result = await compile(source); // uses PLATEX_SERVICE_URL

Which engine runs?

The library auto-selects the engine — you don’t configure this directly. It resolves serviceUrl from the explicit option, falling back to PLATEX_SERVICE_URL.

Mode	When	Engine	Use case
Remote	serviceUrl resolves	Tectonic (on the service)	Production on Vercel / any edge runtime
Local	No serviceUrl, system TeX found	pdflatex / xelatex / lualatex	Self-hosted or dev with TeX Live
Local fallback	No serviceUrl, no system TeX	Bundled Tectonic binary	Dev / CLI without TeX Live installed

Note

The edge-only entry point (@nandan-varma/platex/client) never compiles locally. If no serviceUrl resolves it throws a clear error instead of trying to spawn a TeX process that can’t exist on that runtime.

Develop without any TeX installed

Skip the service entirely and let the library use the bundled Tectonic binary directly — this is exactly what powers the CLI’s zero-install experience:

node scripts/download-tectonic.mjs

The script downloads Tectonic for your platform (macOS or Linux) and verifies it against a pinned SHA256 before extracting. After that, any compile() call with no serviceUrl resolves the bundled binary automatically.

// No serviceUrl → uses the local tectonic binary
const result = await compile(source);

Try the CLI

No project wiring required — the package ships a platex binary that runs the same pipeline:

npx @nandan-varma/platex main.tex     # → main.pdf, no TeX installation required

CLI & watch modeWatch mode, extra files, remote compilation, stdin, and JSON output.

Next steps

Quick startFrom install to your first PDF, end to end.

How it worksDeployment modes, entry points, and what Tectonic is.