PDF Forensics API Pricing in 2026: A Transparent Breakdown

Document fraud detection APIs are priced in three different ways, and the headline number rarely tells you what a check will actually cost you. This article is a snapshot of the market as of October 2026: how the pricing models work, what to compare besides price, and a worked cost-per-check calculation using HTPBE?’s published plans.

Every figure below was read from the vendor’s public pricing page when this was written. Where a vendor publishes no price, the article says so and does not guess.

---

The Three Pricing Models

1. Per-check, with a monthly minimum. You pay a rate per document, but only above a floor. Bynn’s published Document Fraud Detection tiers are one example: $0.79 per check with a $99 monthly minimum, $0.45 with a $299 minimum, and $0.25 with a $599 minimum. The per-check rate looks low, but the effective rate is the minimum divided by what you actually send. At 100 documents a month, the $99 tier costs $0.99 per check, not $0.79.

2. Subscription with an included quota. You pay a flat monthly fee for a fixed number of checks. HTPBE?, Tamperlens and VerifyPDF all work this way, with different quotas and prices. The effective rate depends on how much of the quota you use: a plan that is cheap per check at full utilisation is expensive per check if you use a third of it.

3. Quote-based enterprise contracts. No public price list; you talk to sales. Inscribe’s pricing page says its pricing is based on document volume and features, and directs visitors to book a demo for a quote. Fortiro’s page says pricing is based on features and number of documents, that products are sold on an annual subscription, and that a sales contact follows a form submission. Neither publishes a rate.

None of these models is better in itself. Quote-based pricing fits buyers who need custom volume, data residency or contractual terms. Published pricing fits buyers who want to validate the use case before involving procurement.

---

What to Compare Beyond Price

A lower number per check is only a bargain if the checks are comparable. Before comparing prices, compare these.

Output type. Does the API return a verdict you can route on, a list of signals you must interpret, or a score? HTPBE? returns one of three verdicts – intact, modified or inconclusive – plus named markers explaining the result, and no numeric risk score. A verdict maps directly onto a workflow branch. A bare signal list means you build and maintain the decision logic yourself.

What is being analysed. Some products analyse the file itself (its structure and metadata). Others read the content: extracting figures, matching templates, or checking identity documents. These are different jobs. HTPBE? does the first. It does not read the text of a document, so it cannot tell you that an income figure is implausible; for that you need a data-extraction product alongside it.

Latency. The API reference on GitHub lists typical response times of 2–5 seconds for average PDFs and 5–15 seconds for large, complex ones. If checks run inside a synchronous applicant flow, this matters more than a few cents per check.

Integration effort. Look for a documented REST endpoint, a test mode, and a clear error contract. HTPBE? accepts a public URL to the PDF (up to 10 MB) and returns the result in the response. Every account includes htpbe_test_ keys that return deterministic results without touching your quota, so the integration and all three verdict branches can be built and tested before any live spend.

Data handling. Where do documents go, and for how long? If documents cannot leave your network, ask whether the vendor offers a self-hosted deployment and whether it is priced publicly. HTPBE?’s Enterprise tier is an on-premise deployment, and it is quote-based.

Billing terms. Monthly or annual commitment, whether unused quota rolls over, and what happens when you exceed the quota. Compare these on the same basis – an annual-billing rate is not comparable to a month-to-month rate.

---

HTPBE?’s Published Plans

Plans are month-to-month. Unused subscription checks do not roll over. For occasional volume there are also one-time check packs that never expire while the account is active: $5 for 8 checks and $15 for 25 checks. New accounts receive 5 welcome credits. Web uploads and API calls draw from the same pool. The pricing page is the source of record and shows the current figures in your local currency.

---

Worked Example: Cost Per Check

The cost per check is the monthly price divided by the checks you actually run.

Full utilisation of each plan:

A team that runs 300 documents a month. Growth costs $149, so the effective rate is $149 / 300 = $0.50 per check – the quota is not fully used. Starter cannot cover it, and Pro would work out to $499 / 300 = $1.66 per check. Growth is the right fit.

A team that runs 400 documents in one unusually busy month. This is above Growth’s 350. Moving to Pro at $499 gives $1.25 per check for that month. Covering the extra 50 checks with two $15 packs instead costs $149 + $30 = $179, about $0.45 per check, without changing plan. Packs are spent after the subscription quota, so they work as a buffer for irregular peaks.

A team that runs 1,200 documents every month. For steady volume at this level, Pro at $499 / 1,200 = $0.42 per check is the plan built for it.

The general point: compare plans by price divided by the checks you will really run, and use packs for occasional spikes, not as a substitute for a plan that matches your steady volume.

---

How Published Competitor Prices Compare

Per-check numbers across vendors are published units, not measures of equivalent work. One vendor’s ‘check’ may cover a different set of analyses than another’s, so a lower number does not mean a lower cost for the same detection. With that caveat, these are the self-serve prices found on vendors’ own pages in October 2026:

Read plainly: on published per-check price alone, HTPBE? is not the cheapest option in this market. Tamperlens and VerifyPDF list lower prices per unit. The reasons to choose a product are the output you get, how it fits your workflow, and how you want your documents handled – the comparison pages set out those differences vendor by vendor. Do your own test on a sample of your real documents before deciding.

---

What HTPBE? Does Not Do

• It does not read document content. It cannot check whether the numbers on a bank statement add up or match a tax return.
• It does not establish authenticity. It reports whether a PDF shows evidence of modification. A document created from scratch in professional software can come back intact, because nothing was modified after creation. See PDF forensics without the original file for the limits of this approach.
• It does not perform identity checks. Face matching, liveness and ID template validation are outside scope; this comparison explains how the two layers fit together.

The inconclusive verdict is a routing signal rather than a failure: it means the file does not carry the traits of institutional software. For a document that should have come from a bank, payroll system or tax authority, that is worth a manual look.

---

Choosing a Tier

• Building the integration: start with the htpbe_test_ keys, which are free and unmetered, then Starter for live validation.
• Steady monthly volume: divide each plan’s price by the checks you will really run and pick the lowest effective rate whose quota covers you, as in the examples above.
• Documents that must stay inside your network, or volume beyond Pro: Enterprise, via the API page.

Before choosing a tier, the forensic PDF API page shows what a single check examines and returns.

Sign up to get test keys and a live key. The API reference documents every endpoint, field and error code.