PDFEndpoint Pricing and PDFEndpoint Alternative with Yearly Billing and a Dated Archive of Every PDF

We would rather you pick the right tool than the one whose page you happen to be reading. PDFEndpoint's ladder is one of the lowest in the HTML to PDF API market at small volumes. A side project or an internal tool that makes 800 PDFs a month pays $14 on PDFEndpoint and $49 on our monthly Starter API. If you only need a file back and will never be asked for it again, that $35 a month is real money and PDFEndpoint is the sensible choice.

The picture changes in three places. The first is yearly billing: PDFEndpoint does not publish a yearly price, while SitePDF charges half the monthly rate when you pay for a year, so 10,000 PDFs cost $24 a month instead of $39. The second is the steps between plans. At 50,000 a month PDFEndpoint asks $129 against our $99, and at 200,000 it asks $429 against our $299. The third is the month you overshoot the top plan. A 260,000 PDF month on PDFEndpoint's 200,000 plan is $429 plus 60,000 conversions at $0.01, or $1,029.

For the rest of the field at the same volumes, including PDFShift, DocRaptor and Urlbox, see our PDF API pricing comparison.

PDFEndpoint bills per request, not per page, so a 40 page report and a one line receipt each use one conversion. API responses report "billed_units": 1. Requests sent with sandbox=true are free and carry a watermark, which is how you develop without spending the allowance.

On failed requests the documents disagree with each other. The pricing page says "We count only successful requests as conversions, so you won't be charged for failed conversions." The Terms and the Service Agreement, both version 1.1 dated 28 January 2025, say that "failed requests that fail due to incorrect request data (such as invalid HTML, malformed parameters, or other client-side errors) count towards Your monthly limit", and that only failures on PDFEndpoint's own infrastructure are free. The contract is what governs a billing dispute, so budget for retries caused by your own bad input.

Unused conversions do not roll over. Usage emails go out at 75 and 100 percent of the limit (the pricing page adds 90 percent). Upgrades are prorated and downgrades take effect at the end of the billing period.

On a paid plan nothing stops. PDFEndpoint keeps converting and bills the extra conversions on the next invoice, at $0.04 each on the 500 and 1,000 plans, $0.03 on 2,500 and 5,000, $0.02 on 10,000 and 25,000, and $0.01 from 50,000 up. You can switch overage off or set a ceiling in the account, and the API then answers with error PE-L004 (limit reached, overage off) or PE-L005 (overage ceiling reached). The free plan simply stops until the next period.

That is a fair design, but it puts the decision on you. With overage off, a busy month ends in failed renders for your customers; with it on and no ceiling, the overage rate on the small plans is ten times the per PDF price of the plan itself. Watch the jump from the 5,000 plan: 9,000 conversions there cost $24 plus 4,000 at $0.03, $144, while the 10,000 plan is $39.

By default PDFEndpoint returns JSON with a link on storage.pdfendpoint.com and an expires_after timestamp. It can also return base64 or the raw PDF, call a webhook, or upload to your own Amazon S3 or Google Cloud Storage bucket. Its data processing agreement (16 April 2026) says customer data is processed transiently and "not retained after the conversion is complete, except as necessary for temporary caching". How long a hosted link lives is not published.

For a lot of teams that is fine. For invoices, statements, policy documents and anything else a customer, auditor or court may ask about later, it means building the record yourself: a bucket, a retention rule, a way to find the file by customer and date, and proof of what the page looked like when it was rendered. SitePDF keeps a dated archive copy of every render on every plan, 90 days on Starter, 1 year on Growth, 3 years on Scale and 7 years on Enterprise, retrievable by ID from the API. Our feature overview shows what the archive record holds.

PDFEndpoint's homepage says the service "is backed by a 99.99% monthly uptime guarantee". The Service Agreement is more modest: it commits to commercially reasonable efforts, says it does not guarantee uninterrupted service, and offers no service credits. Its one remedy is that if the service is down for more than three consecutive hours more than three times in a calendar week, you may terminate with a pro rata refund. There is a public status page on UptimeRobot.

PDFEndpoint is run by a small Austrian company, processes data mainly in the EU with AWS, Google Cloud and Vercel as subprocessors, publishes a GDPR DPA with EU standard contractual clauses, and does not claim SOC 2 or ISO 27001. If your security review asks for those, raise it with them before you build. Its docs carry a 2023 copyright and the newest blog post is from August 2024, while the legal pages were revised in 2025 and 2026.

The middle row is the common case for a US business that renders customer documents. The storage line is not just a bucket fee; it is the code that names, indexes and expires files, and the day someone has to find statement 18,442 from last March. If you are still sizing the work, our guide on what an HTML to PDF API really costs walks through the hidden lines.

Both APIs take a URL or a full HTML document and render it in a real browser, so templates move without changes. The parameters map closely: url and html stay as they are, page_size becomes format, wait_for_selector becomes wait_for, and header_html and footer_html become header_template and footer_template. If you use PDFEndpoint's Liquid templating, render the Liquid in your own app first and send the HTML. If you upload to your own S3 bucket today, you can keep doing that with the returned file, and the archive copy is kept either way.

Run both for a week on the same documents and compare the files before you switch. Other vendors buyers usually weigh at this stage are covered in our PDFShift pricing comparison and Urlbox pricing comparison.