Two requests sound identical and cost differently
A web application and API pentest starts at $5,000. The final price depends on the number of user roles, features, API methods and environments, and is fixed in a written proposal before work starts. To see why, picture two requests that sound identical: “check our web application”.
The first has three user roles and one payment flow. The second has seven roles, an API open to partners and a separate test environment. They are priced differently because of the number of places where someone has to check who is allowed to do what.
Four parameters that make up the scope
The scope of a pentest follows the number of places where the application decides who may reach data and actions, and the places where it trusts input. The four parameters named on the service page describe exactly that.
A hypothetical example. Say a SaaS has three roles and forty API methods. That is 120 role-and-method pairs, and for each one someone must answer whether access should be open or closed. Add a fourth role and there are 160. The numbers are made up. They only show how quickly the table grows.
| Parameter | Why it changes the scope | What to prepare |
|---|---|---|
| User roles | Every role adds checks: what it can do, what it must not be able to do and whether it can see other customers’ data. | A list of roles and a test account for each |
| Features | Features that handle permissions, payments or data uploads need their own test scenarios. | A list of key features and the ones you consider riskiest |
| API methods | Each method is checked twice: does the API know who is calling it (authentication), and is that caller allowed this specific action (authorization). Methods in scope are checked this way under the relevant roles. | Documentation or an OpenAPI specification |
| Environments | The environment determines which actions are acceptable. A test environment and a production system need different limits. | An agreed test environment and, for production, availability requirements |
What the work includes
First we agree the application, API, environments, accounts and testing boundaries. Then I review authentication, authorization, input handling and relevant business logic, and afterwards I validate attack paths and their impact within the agreed scope.
The work is manual first. It is guided by the OWASP Web Security Testing Guide and ASVS, and scanners only support coverage. The OWASP project describes ASVS as a framework of security requirements for designing, developing and testing modern web applications and web services. Every finding is confirmed before it is reported.
A scanner sees the request. It does not see that record 1043 belongs to another customer: ownership is recorded in your database, not in the request. That is why manual testing matters for access flaws in particular. OWASP ranks Broken Access Control first in the Top 10:2025, and object-level authorization failures open the 2023 API Security Top 10 as API1. Both categories centre on one question: does the application check that a record belongs to the person asking for it?
What you receive
The result has three parts. First, confirmed findings with evidence, reproduction steps and a severity rationale. Second, fix guidance connected to the affected feature or code. Third, a technical walkthrough of priorities, limitations and next steps.
The word “confirmed” matters here. Your developer does not get a line from a scanner. They get a sequence of requests that shows what happened and an explanation of why it matters for your product specifically.
The report describes what was checked within the agreed scope. It does not prove that everything else in the product is safe, and it states its limitations plainly. A report that promises more than it checked is worse than none: without a report the team knows it was not tested, and with such a report it treats an area as closed that nobody opened.
What is scoped separately
Retesting is agreed for each engagement in the proposal: which findings, which version and in what timeframe. Neighbouring work has its own pages. If you only need an API tested, there is a separate page. The question “where is the root cause in the code” belongs to secure code review. If you already have a report and need to close the findings, that is remediation and fix verification, where developer consultation and verification of fixes can be ordered separately.
I compared code review and pentesting in the article “Security code review or penetration testing: what should you buy?”. If you are unsure, start there.
A starting price marks a lower bound. I name the exact boundaries of the work in the proposal, because without a list of roles, methods and environments any figure would be a guess. If the number of roles, methods and environments turns out larger than in your first description, the total changes, and you see that in the proposal rather than after the work.
How to prepare your request
To start, I need a description of the application and API, documentation or an OpenAPI specification, test accounts for each role and an agreed environment for testing.
Beyond that, I suggest writing down three things in advance. First, a list of roles and what each is forbidden to do. Second, two or three scenarios where a mistake would cost you the most, such as other customers’ data, money or administrative actions. Third, availability requirements if a production system has to be tested: when load is unacceptable and which actions are off limits.
Please do not send production passwords by email. The assessment needs test accounts, not your own.
How to get the exact price
Describe the product, the stack and the reason for the assessment: a release, a significant change or a customer request. I reply within three days, either with questions about the product or with an outline of the proposal. The exact price appears in a written proposal before work starts.
The service description, scope and answers to common questions are on the page Web & API penetration testing. You can send a request through the contact page.
Sources
Need this checked in your own product?
Web & API penetration testingWant to learn to find vulnerabilities like this yourself?
Web application penetration testing course, one to one