Welcome to Tech Spotlight. I’m your host, bringing you breakthroughs in developer creativity. Today we’re exploring how Erik Steiger built a PDF rendering pipeline capable of “generating 1 million PDFs in 10 minutes” for under 35 cents. Faced with legacy systems that took minutes per document, he thought, “this can be implemented so much more efficiently,” and turned to Rust, AWS Lambda, SQS and S3.

Steiger replaced slow tools like Puppeteer and LaTeX with Typst—the “very promising new typesetting system”—and wrapped it in his Papermake library. He designed two Rust Lambdas: one to queue render jobs, the other to fetch templates, render PDFs in about 35 milliseconds apiece and upload results to S3. Using caching for templates and “world caching” for compiled layout, he slashed rendering latency.

All infrastructure was defined in Terraform—API Gateway entry point, SQS queue, Lambda functions on ARM64, and S3 buckets. With batching, concurrency tuning and provisioned scaling, he achieved 91 PDFs per second under default limits and estimates that 60 concurrent workers could reach the 1,667 PDFs per second target.

For the full guide and source code, check out Papermake-AWS on GitHub and start scaling your PDF workloads today.
Link to Article