PDF SDK Pricing and Licensing: What 10 Vendors Actually Charge

We re-read all ten vendor pages on 3 September 2026, seventeen days after the original survey. One vendor moved, and it moved in the direction almost nobody in this category moves. The null results are published here too, because in a market where half the vendors quote rather than publish, knowing that a price did not change is worth as much as knowing that it did.

Syncfusion restored its published pricing. This is the real change. On 17 August every tier on the Essential Studio pricing page read Get a Custom Quote. On 3 September the page publishes per developer figures again: the Document SDK at $1,199 per developer per year with a stated five developer minimum, shown as $5,995 per year for 5 developers; the PDF Viewer SDK at $599 per developer per year, or $2,995 for five; and the UI Component Suite also at $1,199. That takes the count of vendors publishing a full rate card from three to four, and it is the only case we have recorded in this category of a vendor going back to transparent pricing after withdrawing it.

IronPDF: unchanged, including the discount. List prices remain Lite $999, Plus $1,499, Professional $2,999 and Unlimited $5,999, and Professional and Unlimited were again showing as $2,399 and $4,799 on the day we read them. A discount that has now been present on two separate readings seventeen days apart is worth treating as the working price rather than a promotion, though Iron Software does present it as a markdown.

QuestPDF: unchanged. Still $1,999 Professional and $4,999 Enterprise plus local tax, still free under $1M in annual gross revenue, and the exclusion language is still verbatim "government and public-sector organisations, except qualifying public academic institutions, and publicly traded companies".

Apryse: unchanged. Still exactly one published figure, "entry-level pricing for simple use cases starting as low as $1,500", with everything above it behind a sales conversation.

Datalogics: unchanged. The Adobe PDF Library still reads "Starting at $5,999/ year", with the model varying between OEM, SaaS and end user.

Nutrient, iText and ComPDF: unchanged, and still publishing nothing. Nutrient states "To get a price, submit a request through the Contact Sales form." iText still describes its OEM Distribution License as "customized and is calculated depending on your specific use case" and its volume subscription as "based upon the quantity of PDF files you process". ComPDF still says "Our pricing depends on your unique needs: products, features, and deployment options", now alongside an offer of "200+ free API calls per month" through its portal.

Foxit: unchanged in substance, but its pricing URL is gone. The address we used in August, foxit.com/pdf-sdk/pricing.html, now returns a 404, and the SDK has moved to developers.foxit.com, which publishes no figure and offers "REQUEST A QUOTE" and a 30 day trial. We note the 404 because a dead pricing URL is easy to misread as a vendor that stopped publishing, when in this case there was never a published price to lose.

The practical lesson from a seventeen day gap is that the two things most likely to be stale in any PDF SDK comparison are promotional pricing and URLs, not the underlying license models, which barely move. Check the date on any figure you are handed, including ours.

For a single platform, one product, and no redistribution rights, the published numbers land between $999 and $5,999. QuestPDF asks $1,999 for a perpetual Professional license covering unlimited developers in one legal entity. IronPDF starts at $999 for a Lite perpetual license and reaches $5,999 for Unlimited. Apryse says entry level packages start at $1,500. Datalogics says the Adobe PDF Library starts at $5,999 a year. Those four figures are the only real anchors published in the whole category.

Where costs actually escape that range is redistribution. If you are shipping the SDK inside software other people buy, whether that is a desktop product, an on premise install or a SaaS platform, the license you need is not the one on the front of the pricing page. IronPDF charges $1,999 on top for royalty free redistribution. Datalogics describes OEM as a flat annual fee per platform plus royalties based on distribution, which is an open ended term you negotiate. Nutrient offers OEM licensing as unlimited deployment under a single agreement, priced privately. QuestPDF is the outlier here: royalty free redistribution is included at every paid tier, and end users need no license of their own.

The second escape route is platform count. Almost every vendor in this list prices per platform. A team that needs Web, iOS, Android and a server side pipeline is buying four things, not one, and the quote reflects that. This is the single biggest reason published starting prices under estimate real invoices, and it is why the quote only vendors can defend not publishing a number: there genuinely is no one number.

Because they price on the value your product delivers rather than on the cost of the software, and that is easier to do when the buyer arrives without an anchor. Nutrient states the logic almost outright on its own pricing page: pricing reflects the components licensed, the deployment model, and the scale and nature of your use case, and the sales team builds a quote tailored to how you are using the technology and the value it provides. That is value based pricing described plainly, and it is a legitimate model. It also means a well funded fintech and a two person agency will not be quoted the same figure for the same binary.

The structural reasons are real too. A component based SDK with six platforms and a dozen add on modules has hundreds of valid configurations. OEM and royalty terms depend on your distribution volume. Air gapped and regulated deployments carry support obligations that a self serve checkout cannot capture. And in enterprise sales, a published list price becomes a ceiling you can only discount from.

What buyers should take from this is not outrage but planning. If a vendor publishes no price, you are in a procurement process, not a purchase. Build in three to six weeks, get the redistribution and platform scope written into the first quote rather than the second, and ask for the renewal terms in writing before you sign. Syncfusion moving to Get a Custom Quote on every tier and then back to published per developer pricing inside three weeks is a reminder that even the published numbers in this category are not permanent.

The dollar figure matters less than the unit it is attached to. Five models are in use across the vendors above, and mixing them up is how a budget gets blown at renewal.

• Per legal entity, unlimited developers. One price covers everyone in the company. QuestPDF works this way now, and explicitly dropped its earlier developer count limit. Predictable, and it stops seat counting.
• Per developer and per location. The license scales with headcount and with the cities your team sits in. IronPDF prices this way, with location counts of one, three, ten and unlimited across its tiers. Cheap for a small team, expensive for a distributed one.
• Per platform, component based. You license a base capability and add modules. Apryse and Nutrient both work this way. Flexible, and very hard to forecast until someone scopes it with you.
• OEM with royalties. A flat annual fee plus a per unit or per distribution royalty. Datalogics describes exactly this for the Adobe PDF Library. This is the model where costs grow with your success, and the one to model hardest before signing.
• Per server or per desktop, end user licensing. Priced by deployed instance. Common for internal enterprise tooling, and a poor fit for anything that autoscales.

Then there is the perpetual versus subscription split. QuestPDF and IronPDF sell perpetual licenses: you keep the versions released during your paid term forever, and renewal buys updates and support rather than the right to run the software. Nutrient, iText, Datalogics and Syncfusion sell annual terms, where lapsing means you lose the right to ship. For a product with a ten year life, that difference compounds into more money than the sticker price ever will.

Three vendors here offer a genuinely free production license, and all three attach conditions that disqualify a lot of the companies that assume they qualify.

QuestPDF Community is free for organisations under $1,000,000 annual gross revenue, with no registration, key, watermark or time limit, and full commercial use permitted. The exclusions are the interesting part: government and public sector organisations, and publicly traded companies, are not eligible regardless of revenue. Charities and non profit or public academic institutions are eligible regardless of revenue. If your revenue crosses the line, you get a 90 day transition period starting at the end of that fiscal year, with no back payments. For agency and contract work, eligibility follows the client, not the developer.

Syncfusion Community is stricter on every axis. You need less than $1M in annual revenue, up to 5 developers, and no more than 10 employees, and all three at once. There is now a funding clause as well: if you have taken private equity or venture capital money, you must have received less than $3M to remain eligible. A seed stage startup with a $4M round and four engineers does not qualify, which surprises people.

iText AGPL is not a free tier in any practical sense for commercial software. The AGPL obliges you to release the source of your own application under the same license, including for network use. Teams that adopt iText on the free path and later ship a closed source product are the most common licensing accident in this category. The safe free alternatives with no strings are PDFsharp under MIT and Apache PDFBox under Apache 2.0, neither of which renders HTML. Our Java PDF library comparison and C# PDF library comparison go through what each license permits in a closed source product.

It depends almost entirely on volume and on what else you need, and the honest answer is that an SDK wins more often than API vendors like to admit.

An SDK is a fixed cost. Once you have paid the license, rendering a million documents costs you the servers and nothing more. A rendering API is a variable cost that scales with usage. At $1,999 a year for QuestPDF, the SDK breaks even against a typical API rate card somewhere in the low hundreds of thousands of documents a year, and everything above that line is free. If you have high, steady, predictable volume and an engineering team comfortable running Chromium or a layout engine in production, buy the SDK.

The API wins on three things. First, low or spiky volume: paying $1,999 up front to render 3,000 documents a year is poor arithmetic. Second, fidelity to a real browser. Most .NET and Java PDF libraries either do not render HTML at all, or render it with a partial CSS engine that will not match Chrome on modern layout. Third, the operational load you avoid: font packages, headless browser upgrades, memory limits, timeout handling and the render fleet itself.

Our own position, stated plainly because this page is useless otherwise: Sitepdf is a rendering API, planned at $29, $99 and $299 a month for 10,000, 50,000 and 250,000 renders, and it stores a timestamped archive of every page it renders in the same call. If your requirement is high volume HTML rendering and nothing else, a licensed SDK on your own hardware will beat us on unit cost eventually. If your requirement is browser accurate output, no infrastructure, and a retrievable record of what was rendered and when, that is the trade we are built for. Our normalized PDF API pricing comparison works out what ten thousand documents a month costs across twelve rendering services if you want the API side priced properly.

Five steps, in this order, based on what the ten vendors above actually require of a buyer.

1. Write down your deployment shape first. Platforms, whether you redistribute, whether any environment is air gapped, and whether you are one legal entity or several. Every quote in this category is built from those four answers, and having them ready removes a week of back and forth.
2. Test rendering fidelity before you talk to anyone. Take the ugliest real document you have, not a sample invoice, and run it through the trial. Flexbox, grid, web fonts, page breaks inside tables and headers repeating across pages are where the differences show. Note that a Nutrient trial watermarks rendered PDFs, so judge layout rather than appearance there.
3. Ask for the redistribution term in the first quote. Not the second. OEM and royalty language is where the number moves most, and discovering it late resets the negotiation.
4. Ask what renewal looks like in year three. Perpetual versus annual, whether the price is locked, and what happens if you lapse. QuestPDF publishes a price lock; most vendors do not publish anything.
5. Price the API alternative in parallel. Even if you buy the SDK, a rate card from a service that publishes one gives you a number to negotiate against, which is worth more than any listicle.

One last thing worth saying out loud. Every figure on this page has a date attached because in this category that is the only responsible way to publish one. If you find a comparison that lists per developer prices for Nutrient, Foxit or iText, ask where those came from, because as of 3 September 2026 none of those three vendors publishes one.