Leafwright

Leafwright is a platform for generating PDFs, built for teams that need finished documents out of an API without standing up and babysitting their own browser infrastructure. It combines a hosted render API with a visual studio, so you can create documents by hand in a workspace or produce them programmatically at scale, and both sides share the same rendering engine so what you design is what your API returns.

Anyone who has built PDF export into a product knows the usual path. You spin up headless Chrome, wire it to a queue, fight with fonts and page breaks, and then own the memory leaks and crashes forever. Leafwright's reason to exist is to take that whole burden off your plate. You send it your content, it renders the PDF, and you never run or maintain a browser cluster yourself. That's the trade it's offering, less infrastructure in exchange for a metered service.

The render API is flexible about what you feed it. You can send raw HTML, point it at a public URL, or call a published template with JSON data, and templates support Handlebars placeholders that get filled from the per-request JSON. That covers the common cases cleanly, from turning an existing web page into a PDF to generating thousands of personalized invoices from a single template. Around the render itself, it gives you project-based API keys with Bearer auth and key rotation, signed delivery URLs so files stay private, webhooks and render events for notifications, automatic retries, and job logging so you can see exactly what happened. Those are the unglamorous pieces that a homegrown setup usually skips and then regrets, since a PDF that silently fails to render or leaks a private file is the kind of bug you only find in production.

The studio is the other half. It lets you draft documents with AI from a text description, edit on a canvas with live preview, and lean on a template system so structured documents stay consistent. There are inline accessibility and print-validation checks that catch problems before you ship, plus review and approval steps so a document can be signed off inside the tool. A brand kit locks fonts, colors, and footers so everything that comes out looks the same, whether a person made it or the API did.

Keeping the studio and the API on one engine is more than a convenience. It means a template you build and preview by hand renders the same way when your code calls it in production, so there are no surprises where the design looked right in the editor and then broke in the automated output. Teams can design and approve a document once, then trust that every programmatic render matches what got signed off, which is exactly the gap that trips up homegrown PDF pipelines.

There's also an angle for AI workflows. Leafwright ships a remote MCP server with OAuth alongside a local stdio package, which means agents in tools like Claude, ChatGPT, Cursor, and Codex can generate PDFs directly. If you're building automation that needs to hand a user a finished document, that's a clean way to wire it in without gluing together your own rendering step at the end.

It fits developers and document teams operating at some scale. Backend engineers who need reliable PDF output, teams churning out proposals and contracts that must stay on-brand, and products that generate receipts or reports for users are the natural audience. The mix of a real editor and a real API also means non-engineers can build and review templates while developers handle the programmatic side, which keeps both groups working inside the same system. That shared workspace is part of the appeal, since design changes don't have to travel through a developer ticket to reach production and a marketer can adjust a footer or a color without touching code.

Pricing is freemium and usage-based. A free Hobby plan gives you 100 PDFs and 8 AI drafts with no credit card, and paid tiers climb from Indie at $19 a month for 3,000 PDFs, to Pro at $79 for 15,000, to Business at $299 for 80,000, with a custom Enterprise tier above that. One credit renders up to a 50-page PDF and longer documents use more credits per 50-page block, paid plans support metered overage through Stripe, and extra AI drafts run between $0.20 and $0.40 each depending on your plan. It's the kind of pricing that stays cheap while you're small and scales with how much you actually render, so a side project can start on the free tier and grow into the paid plans only once the document volume genuinely justifies the added spend.