CloudConvert API Cost for 10,000 HTML to PDF Renders a Month: Where the Bill Actually Comes From

Ten thousand HTML to PDF renders a month on the CloudConvert API costs $74.00 if every render finishes inside a minute, and close to double that if they do not. The document count is the same in both cases. What moves is render time, because CloudConvert bills a base credit per conversion plus one more credit for every minute the conversion runs. Most teams discover which case they are in after they integrate, when the first full month lands. This is how to work it out beforehand, priced on rates read from CloudConvert on 22 September 2026.

The team we are pricing

A US analytics SaaS sends its customers a monthly performance report as a PDF. The report is a page in their own app: server-rendered markup, a web font, four charts drawn by a JavaScript library, and logos pulled from customer accounts. They have 8,000 customers and render roughly 10,000 PDFs a month once you count re-runs, support requests and the two customers who download the same report eleven times.

They are not doing anything exotic. This is the single most common shape of HTML to PDF work: an existing page in an existing app, turned into a file on a schedule. It is also, as it happens, the shape that a time-based meter prices worst.

The good case: $74.00 a month

CloudConvert charges credits. An HTML or website to PDF conversion falls under General, the cheapest base cost on their card at 1 credit. Office to PDF and iWork to PDF cost 2, and PDF to Office costs 4. So if every one of those 10,000 reports renders in under sixty seconds, the job needs 10,000 credits.

Ten thousand credits on a monthly subscription costs $74.00, which works out at $0.0074 a credit, or $7.40 per thousand renders. Bought as a one-time package instead, where credits never expire, the same 10,000 costs $134.00. Those are base prices. CloudConvert displays its prices inclusive of the viewer's local VAT, so a reader in a 23 percent VAT country sees $91.02 where a United States buyer pays $74.00. We wrote down every rung of that ladder in the CloudConvert API pricing breakdown, because the numbers live inside a slider and no page we could find lists them together.

At $74.00 a month, this is a perfectly reasonable bill and there is nothing to fix. The question is whether it is your bill.

The clause that moves it

CloudConvert is direct about the meter: "The longer a conversion takes, the more resources it consumes and the more expensive it becomes. Our packages and subscriptions typically consume one credit per minute of conversion time." And then: "By default, conversions consume at least one credit, with additional credits charged for every extra minute if the conversion takes longer than one minute."

For most of what CloudConvert converts, that is a non-event. They say so: "Ninety-five percent of our conversions take only a few seconds, and we already account for that." Across a mix dominated by images, audio, video and Office files, that is believable. A DOCX is a zip of XML and converting it is deterministic work on a known quantity of bytes.

Rendering a web page is not that. A headless browser has to fetch the document, then fetch every asset the document asks for, execute the JavaScript, wait for the layout to settle and only then paint to PDF. The slow parts are almost never your markup. They are a web font served from a third party, a charting library that renders after an animation frame, images loaded lazily, and an API call the page makes on load that has to come back before the numbers appear. Our analytics SaaS has all four.

The awkward part is that none of this correlates with anything on the rate card. File size does not predict it. Page count does not predict it. Two reports of identical length can differ by a factor of ten because one customer has 40 data points and another has 4,000.

How to find out which case you are in, in about ten minutes

You do not need the vendor for this, and you should not guess. Take the ten heaviest real documents you produce, not the sample invoice in your test suite, and time the render end to end.

If you already render locally with a headless browser, time that. If you are evaluating CloudConvert cold, their free tier gives you 10 credits a day, which is enough to push your worst five templates through and read the job timings back from the API. The number you want is not the average. It is the share of your monthly volume that crosses sixty seconds, because that share is billed twice.

Three things worth knowing while you measure. CloudConvert's free tier caps processing time at 5 minutes, files at 1 GB and concurrency at 5 tasks, all of which become unlimited on any paid plan, so a timeout on the free tier is not necessarily a timeout in production. Failed conversions are not billed, which is genuinely useful when you render pages you do not fully control. And the clock is conversion time on their side, so a slow response from your own API, which the page waits on, lands on your bill.

The bad case, priced

Say the measurement comes back unhappily: the charts and the font push a typical report to about seventy seconds, so most renders cost 2 credits. The 10,000 reports now need roughly 20,000 credits. On CloudConvert's ladder the 19,000 rung is $131.00 a month and the 25,000 rung is $168.00, so the bill lands a little above $131.00.

Nothing about the product changed. The same customers got the same reports. The bill went from $74.00 to about $131.00 because a font and a chart library took an extra ten seconds. That is the kind of line item that quietly compounds across a year and never gets questioned, and it is exactly what a SaaS spend review is built to surface: a read-only view of what every tool is actually costing tends to find two or three of these before it finds anything dramatic.

Three fixes, and what each one costs

Make the page faster. This is the fix nobody wants and it is often the right one, because it pays off everywhere, not just on the PDF bill. Self-host the web font instead of fetching it, render charts server-side or as static SVG, inline the logos, and make sure the page does not wait on a live API call it could have been handed as data. A day of engineering time, and the reward is that you drop back under the one minute line and the bill halves. It also makes the report faster for the customers reading it in a browser.

Buy more credits. The honest option. If 20,000 credits at about $131.00 a month is affordable and the engineering time is not, buy the credits and move on. At this scale the difference is $684 a year, which is less than a day of most people's engineering time. Do the arithmetic before you decide it is a problem.

Change what is being metered. If your render times are unpredictable and you would rather not own that variable, a per-render tier prices the document rather than the compute. Our Starter API covers 10,000 renders a month at $29.00, or $174.00 a year, and the slow report costs the same as the fast one because nobody is holding a stopwatch. We sell that, so weigh it accordingly, and read the next section before you act on it.

When staying on CloudConvert is the right answer

If the PDF is one job among several, stay. CloudConvert converts roughly two hundred formats, and the moment your application also has to accept a customer's DOCX, transcode a video or normalize an image, a single-format rendering API is the wrong tool and running two vendors to save $45 a month is a false economy.

Stay too if your volume is seasonal. Package credits never expire, which suits a business that renders almost nothing for nine months and then produces a quarter of a million documents in one quarter better than any monthly tier does, ours included. And stay if you burst hard: CloudConvert gives paid plans unlimited concurrent tasks, where our tiers cap concurrency at 3, 10 and 30 renders. If you generate fifty thousand statements in a single overnight window, that ceiling is a real constraint you would have to design around.

What actually decides it is not price, it is whether you want to own the render-time variable. If your documents are uniform and quick, credits are good value and the meter never bothers you. If your slowest template is ten times your fastest and the mix changes as customers grow, a flat count removes a number you cannot control. For teams weighing the wider field rather than this one decision, the CloudConvert alternative comparison prices ConvertAPI, Zamzar, PDF.co and Adobe PDF Services on the same basis, and the PDF API pricing comparison puts twelve vendors on one 10,000 document workload.

What to do this week

Time your ten worst templates and write down what share crosses sixty seconds. Multiply your monthly volume by 1 credit, then by that share again, to get your real credit count. Price that number on the ladder rather than on the entry tier, because the entry tier is almost never where you land. Then decide whether the gap is worth a day of engineering, more credits, or a different meter.

If you want to see what a render looks like before any of that, the HTML to PDF converter will turn a document into a PDF in the browser without an account, and the HTML to PDF API roundup covers who else is in this market. Our own rates are on the pricing page.