Sitepdf · HTML to PDF + Archive API
The journal
Long-form, practical pieces on HTML to PDF conversion, page capture and archives that hold up. No launch hype, no invented case studies.
9 September 2026 · 8 min read
Browserbase prices its plans around agent tasks that run for minutes. A PDF render takes about four seconds. That mismatch makes Browserbase far cheaper for document work than its own marketing suggests, which means price is almost certainly not your reason for leaving.
Read the piece8 September 2026 · 8 min read
Bannerbear prices images clearly and refuses to price PDFs at all, saying only that they "use more credits". For document workloads that gap is the whole decision. Here is what four competitors publish instead, the second meter hiding in Bannerbear plans, and how to tell whether switching pays back.
Read the piece6 September 2026 · 8 min read
Every guide to Playwright on Lambda stops at the per second rate. The figure that actually moves the bill is what a hung invocation costs, because AWS charges the timeout rather than the render.
Read the piece2 September 2026 · 8 min read
Dropbox Sign gates embedded signing behind a $250 a month plan when its entry API tier is $75. Documenso charges $40 to embed and $250 to remove its branding. Two vendors charge nothing extra at all. The tier, not the per signature rate, is what a product integration actually pays.
Read the piece4 September 2026 · 8 min read
Adobe publishes one number for the PDF Services API and it is not a price. The 500 free Document Transactions buy 25,000 pages on a Create call and 50 pages on an Accessibility Auto-Tag call. Here is the endpoint by endpoint table, three worked workloads, and what the paid tier does not say.
Read the piece3 September 2026 · 8 min read
A compliance letter is not a campaign with a smaller list. It goes to one person, at one moment, usually first class or certified, and someone may later have to prove it was sent. Here is what four print and mail APIs charge for certified mail and return receipts, why two of those numbers cannot be compared to the other two, and which vendors are built for a trigger rather than a send date.
Read the piece2 September 2026 · 8 min read
All three sell credits rather than requests, and the plan prices look like a coin flip until you divide. Here is the cost per 1,000 JavaScript rendered pages at every tier, the anti-bot multipliers that move the bill five to fifteen times, and the one vendor whose feature costs are not published at all.
Read the piece24 August 2026 · 8 min read
Aspose metered licensing starts at $1,999 a month and bills a credit every time you open or save a document, which makes a straight conversion two credits rather than one. Here is what Aspose actually publishes about the unit, the 20MB surcharge nobody models, and the point where the $1,199 perpetual license is simply the cheaper purchase.
Read the piece22 August 2026 · 9 min read
Two of eight vendors in this market publish a rate card. Here is every figure we could read ourselves, what each tier actually meters you on, and the second cost that nobody quotes until you are already in the implementation call.
Read the piece17 August 2026 · 8 min read
The Syncfusion Community License has four conditions, not one, and the funding test catches funded startups that assume revenue is the only bar. Here is the exact wording, what happens when you outgrow it, and how the free tiers from QuestPDF, iText, PDFsharp and PDFBox differ in what they let you ship.
Read the piece15 August 2026 · 8 min read
Both categories buy the same keywords and describe themselves with the same three words, and buying the wrong one surfaces six weeks into implementation. Automation is a generator, management is a filing system. Here is the test that separates them, plus what each actually costs.
Read the piece14 August 2026 · 8 min read
Adobe Create operation turns HTML into PDF from a zip archive, a public HTTPS URL or a dynamic template with JSON data. Here is how each path works, the three constraints that catch teams out, and how the Document Transaction maths lands once you go past the free 500 a month.
Read the piece12 August 2026 · 8 min read
Competitor monitoring usually starts with someone checking a pricing page by hand every few weeks and stops when they get busy. Here is how to pick the handful of pages that actually matter, what interval each one deserves, which tools do the watching, and why the copy you keep afterwards is worth more than the alert that told you.
Read the piece11 August 2026 · 8 min read
Apache PDFBox has no HTML parser and no CSS engine, so it cannot convert HTML to PDF, and no version of it will. Here is what PDFBox is genuinely good at, why the gap exists, and which Java library to reach for when your document starts as markup.
Read the piece9 August 2026 · 9 min read
A .NET PDF library is one of the few dependencies where the license, not the feature list, decides whether you can ship. iText is AGPL unless you buy a quote-only commercial license. QuestPDF is free under 1M USD revenue, then 1,999 USD. IronPDF starts at 999 USD. PDFsharp is MIT with no conditions at all. Here is what each one actually costs a closed source commercial application, read from the vendors own sites on 9 August 2026.
Read the piece3 August 2026 · 8 min read
Your Django template looks perfect in the browser and the PDF comes out with a blank rectangle where the chart should be. That is not a bug and no configuration flag will fix it: WeasyPrint runs no JavaScript at all, deliberately. Here is what the documentation actually says, why the decision was made, the four ways to get a chart into the PDF anyway, and the point at which you should stop fighting it.
Read the piece3 August 2026 · 8 min read
Blank PDFs, text sliced at a page boundary, blurry output, vanishing backgrounds, and an error the moment you try to run it in Node. Almost every html2pdf.js complaint traces back to one design decision: it rasterizes your page into an image rather than converting it. Here is what that causes and what to do about each symptom.
Read the piece1 August 2026 · 9 min read
An HTML to PDF API is almost always a data processor under GDPR, which makes a written DPA, a sub-processor list, a lawful transfer basis and a stated retention period mandatory rather than optional. Here is where the data actually goes, what the Data Privacy Framework status is in August 2026, and the seven questions to put to a vendor before you integrate.
Read the piece31 July 2026 · 9 min read
Your public website is a retail communication, so the rules that cover email cover it too. Retention is three years, the first two in an easily accessible place, and since the 2022 amendments WORM storage is no longer the only compliant format. Here is what the rule text actually says and where firms fail an exam on it.
Read the piece24 July 2026 · 8 min read
Zapier, Make and n8n move data between apps; none of them turn that data into a document. There are three ways to add the missing step, they suit different teams, and the one you pick decides how much of the work lands on an engineer. Here is the honest comparison, plus the four things that break in every route.
Read the piece19 July 2026 · 9 min read
The page looks perfect in Chrome and the PDF comes out looking like 1996: no colors, no layout, everything stacked in one column. CSS loss is the single most reported HTML to PDF problem, and it has five distinct causes with five different fixes. Here is how to tell which one you have hit.
Read the piece19 July 2026 · 8 min read
Your logo is missing, the product photos are empty boxes, and the chart image never appears. Images fail in HTML to PDF for four specific reasons, and lazy loading is the one almost nobody expects. Here is how to diagnose each and the fixes that make images render every time.
Read the piece19 July 2026 · 9 min read
The DOJ ADA Title II rule put PDF accessibility into procurement questionnaires, and generated documents are in scope. Here is what a tagged PDF actually is, what headless Chromium produces, the post-processing step that silently strips it all out, and where automated conversion stops being enough.
Read the piece24 July 2026 · 8 min read
Most teams go shopping for a renderer first. That is the wrong first question. There are four ways a document gets generated automatically, they fail in different ways, and picking the pattern before the vendor is what keeps a document pipeline from quietly dropping files on its first busy day.
Read the piece19 July 2026 · 9 min read
Someone on your team rebuilds the same report by hand every Monday. Automating it takes four moving parts: a scheduler, a report URL with an explicit date range, a render call that waits for the charts, and delivery that stores what it sent. Here is how each should work, and the failures that make automated reports look homemade.
Read the piece19 July 2026 · 9 min read
A PDF renderer is a browser running inside your network, loading HTML that a user influenced. That is the exact shape of a server-side request forgery bug, and PDF generators have been a reliable place to find one for years. Here are the two attacks, the controls that stop them, and the eight questions that decide a vendor security review.
Read the piece19 July 2026 · 10 min read
Comparing HTML to PDF API pricing is harder than it looks, because almost no two vendors bill the same unit. One charges per document, one per half megabyte, one per thirty seconds of browser time. Here is how to translate them all into one number, and the real break-even point against self-hosting.
Read the piece19 July 2026 · 8 min read
You set a brand font, it looks perfect in the browser, and the PDF comes out in Times New Roman. Web fonts are one of the most common HTML to PDF failures, and it is almost always a timing or reachability problem, not a bug. Here is why the font falls back and the three fixes that make it embed every time.
Read the piece19 July 2026 · 9 min read
A 200-row report renders fine, a 20,000-row one times out or kills the browser with an out-of-memory error. Large HTML to PDF jobs fail for a handful of specific reasons, and each has a clean fix. Here is why big pages time out, why the renderer crashes, and how to render heavy reports reliably.
Read the piece19 July 2026 · 8 min read
A header that only shows on page one, a footer that overlaps the content, page numbers that will not appear: margins and running headers are the second most common HTML to PDF headache after page breaks. Here is the CSS that sets margins and the two reliable ways to get a header and footer that repeat on every page.
Read the piece19 July 2026 · 8 min read
A US invoice that comes out A4, a wide financial table crammed into portrait, a page size that changes depending on who opens the file: page size and orientation trip up almost every HTML to PDF setup. Here is how to pin A4 or Letter, switch to landscape for wide content, and why the size sometimes gets ignored.
Read the piece19 July 2026 · 7 min read
You render an invoice with a colored header bar and it comes out plain white, or your branded banner image disappears from the PDF. This is not a bug, it is a print default. Here is why HTML to PDF strips background colors and images, and the single setting that turns them back on.
Read the piece19 July 2026 · 8 min read
A row cut in half, a card sliced across two pages, a heading stranded at the bottom: page breaks are the most common complaint about HTML to PDF. Here is the CSS that controls them, why it sometimes gets ignored, and how to get clean breaks on invoices, reports and tables.
Read the piece19 July 2026 · 8 min read
An empty box where the chart should be is the classic HTML to PDF bug: the PDF was captured before the JavaScript finished. Here is why it happens, the wait signals that actually work (and the one that does not), and how to make dashboards and font-heavy pages render every time.
Read the piece19 July 2026 · 9 min read
Go has no built in HTML to PDF, and half the packages people find are archived. Here is the current map: chromedp for real Chromium output, why gofpdf and the wkhtmltopdf wrappers are dead ends, and when to reach for Gotenberg or a hosted API instead, with code.
Read the piece19 July 2026 · 8 min read
Running headless Chrome in your own app is a real operational burden: memory leaks, cold starts, and hosts that will not let you launch a browser at all. There are exactly two honest ways around it, and they trade off in opposite directions. Here is how to choose.
Read the piece19 July 2026 · 10 min read
Java has three real ways to turn HTML into a PDF: a pure Java engine like OpenHTMLtoPDF, the commercial iText pdfHTML, or a real browser via Playwright or an API. Each has a catch, CSS limits, an AGPL license, or an infrastructure bill. Here is the current map with code.
Read the piece19 July 2026 · 9 min read
The default Rails answer, WickedPDF over wkhtmltopdf, is now a liability: the binary was archived in 2023 with unpatched CVEs and CSS frozen at 2014. Here are the paths that actually work in 2026, Grover, Ferrum, Prawn and a hosted API, with code and when each fits.
Read the piece19 July 2026 · 9 min read
Two very different jobs get called "invoice PDF API": fetching the PDF a billing platform already made, and generating your own from an HTML template. Here is when to use each, with the Stripe and QuickBooks endpoints and the render call, plus how to keep a record every finance team eventually needs.
Read the piece19 July 2026 · 9 min read
Node has no shortage of HTML to PDF packages, and most of them are a thin wrapper around headless Chrome or an abandoned binary. Here is the current map: which ones render modern CSS, what running the browser actually costs, and the working code for each.
Read the piece19 July 2026 · 9 min read
Most PHP HTML to PDF tutorials still hand you dompdf or a wrapper around an abandoned binary. Here is the current map for PHP and Laravel: which library renders modern CSS, which is a security risk, and the code for each.
Read the piece19 July 2026 · 9 min read
Bundling headless Chrome into a Lambda function is a rite of passage that fights you at every step: the deployment size limit, the binary, cold starts, memory. Here is the working setup, and the honest point where it stops being worth it.
Read the piece16 July 2026 · 9 min read
Half the Python HTML to PDF advice online points at a library that shells out to an abandoned binary. Here is the current map: what renders what, the code for each, and where each one falls over.
Read the piece16 July 2026 · 9 min read
The browser can make a PDF and so can your server, and they produce very different files. Where html2pdf.js is genuinely fine, where it quietly ships a screenshot, and when the job moves to Node.
Read the piece14 July 2026 · 8 min read
You already have the invoice as HTML. The job is not designing a PDF, it is rendering the markup you have, at the right moment, and keeping the file you produced.
Read the piece14 July 2026 · 9 min read
Printing is not screenshotting. The renderer swaps your stylesheet, drops your backgrounds, paginates a layout built to scroll, and often fires before your content lands.
Read the piece6 July 2026 · 9 min read
The library is free until you own the headless Chrome fleet behind it. Here is the real bill: memory leaks, font drift, on-call pages, and the point where an API wins.
Read the piece1 July 2026 · 8 min read
Ctrl+P works once. It does not work 4,000 times a month, and it proves nothing about what the page said on March 3. Here is the full toolbox, honestly compared.
Read the piece19 July 2026 · 9 min read
Half the C# HTML to PDF advice still points at a wrapper around an archived binary. Here is the current map for .NET: which library renders modern CSS, which is a security liability, and the code for each.
Read the piece19 July 2026 · 8 min read
Both drive Chromium, so the PDF is the same. The real difference is the API shape, the operational tax, and the moment you should stop running browsers on your own servers at all.
Read the piece1 September 2026 · 9 min read
Six of ten accessibility audit providers publish no price. What a public sector buyer should require in the scope, which VPAT edition to ask for, and the Title II exception that removes most of a document backlog.
Read the piece31 August 2026 · 8 min read
Five alternatives to the Acrobat Pro accessibility checker, priced on 31 August 2026. Every tool that formally validates PDF/UA is free, and exactly one of them runs without a person in front of it.
Read the piece30 August 2026 · 8 min read
Four web accessibility vendors checked on 30 August 2026 and only accessiBe publishes a price. What the published ladder costs, why the other three quote only, and the line item none of the four rates cover.
Read the piece21 August 2026 · 8 min read
Four VPAT providers checked on 21 August 2026 and only two publish a price. Here is what a VPAT and the audit underneath it really cost, why they are billed separately, and the per page remediation trap that catches software vendors whose product generates documents.
Read the piece20 August 2026 · 9 min read
A checker will tell you a clean file is clean. It cannot tell you a clean file is usable. Here is what PAC, veraPDF and the Acrobat checker actually test, the failure conditions no machine can judge, and the five things that break in documents produced by a rendering pipeline.
Read the piece29 August 2026 · 9 min read
Four vendors publish a remediation price and three will not quote without a call. What each costs, and the pre-existing documents exception that removes more from a university backlog than any tool will.
Read the piece24 June 2026 · 10 min read
A screenshot in a shared drive is not a record. For compliance work a snapshot needs a capture method you can explain, a timestamp you can defend, and retention you control.
Read the pieceEarly access
The API opens to the list first, in order. Early access locks the planned launch rates for 12 months. No card required, launching soon.
Render + archive, one API
Early access, no card