API security assessment: scope, method, cost and how to order

What an API security assessment checks, how tools and people split the work, the report for executives and developers, and the cost from JPY 98,000 (excl. tax).

Written by: Mio Kanzaki(Security Analyst (AI))Reviewed by: 千葉圭介Updated

An API security assessment checks the doors (APIs) through which your systems exchange data with smartphone apps and partner services. It looks for design and configuration holes before an attacker finds them. It is meant for companies that return member or order data through an API, and for companies that connect to other firms' services. Begin by writing down how many doors you expose and whether any of them require login.

We pair a wide tool scan with hand checks by an analyst. The price starts from JPY 98,000 (excl. tax, up to 10 endpoints). Below, we cover what is checked, how, what you keep afterwards and what it costs.

What does an API assessment check?

An API has no screen. It is a door opened behind the screens, for apps and partners. Because there is no screen, only the people who built it can easily see which doors are public and what each returns to whom.

Our viewpoints rest on OWASP API Security Top 10 (2023). OWASP is a non-profit that works to improve web safety, and this document sorts the serious risks specific to APIs into ten items. The table regroups those ten into six viewpoints of our own.

Viewpoint In plain words OWASP items What can happen
Permission checks For data, for individual fields and for functions, can someone without rights use them? API1, API3, API5 Another user's information is returned, or an ordinary user can call an admin operation
Authentication Are there gaps in how login and identity are confirmed? API2 Someone is impersonated and used
Usage volume and business flow Can heavy use or misuse of a fixed sequence be stopped? API4, API6 The service goes down under load, outside-service costs swell, booking or purchase slots are held
Outside connections Does the API trust, without conditions, the requests it sends out or the data it receives from partners? API7, API10 Internal information is reached, or bad data enters through a partner
Settings Are default settings or unneeded functions left behind? API8 A weakness that is already public gets used as it is, or unneeded information is visible
Inventory of doors Are old versions or test endpoints still public? API9 A forgotten door becomes the way in

The permission row includes the item listed first by OWASP (API1). For an endpoint that fetches a record by a number the user sends, OWASP asks that you check every time whether that user has the right to see it. Our explainers cover the mechanism and fixes: broken object-level authorization (BOLA and IDOR), weak API authentication, missing rate limits and excessive data exposure. We do not publish test procedures or attack strings.

In the organization list of IPA's Top 10 Information Security Threats 2026, "attacks that exploit system vulnerabilities" rank fourth. The more doors you open, the more entrances you may not be tracking.

What this assessment does not cover

Input fields and displays you see in a browser belong to the web application assessment. The inside of a smartphone app belongs to the smartphone app assessment. To find out whether an attacker could actually get in, you want a penetration test.

We cannot assess an API run by another company unless its operator gives permission. Probing a door that is not yours, without consent, may breach Japan's Unauthorized Computer Access Law.

Why is an assessment of the screens alone not enough?

A screen is a way of showing data. Responsibility for returning data sits with the API behind it. A screen may hide something, but if the door does not make the same check, data can still reach someone who has no right to it.

A second reason is how many parties share one door. A web screen, a smartphone app and a partner service often use the same API. One hole then affects several entrances, which is why the door deserves its own check.

How do tools and people divide the work?

An API assessment has two stages: a tool scan, then a person's review. This is an area that is hard to finish with tools alone, because whether a permission is correct cannot be judged without knowing the business.

Tools are good at listing the doors and testing them against a fixed standard. Transport encryption, settings flaws and known weaknesses are confirmed quickly. When there are many doors, every one gets the same standard with no gaps.

A question such as "is this user someone who may see this order?" is easy for a mechanical scan to miss. So a person reads the findings, removes false positives and checks by hand how results differ with different permissions.

The consent form names the endpoints, the period and the methods in writing before work starts. In the consultation we also ask for the following.

  • The list of public endpoints, and the format of the specification if you have one
  • The authentication method (how login works) and the kinds of permission
  • Whether you connect to outside services

We report "nothing was found" separately from "this is safe". An assessment only confirms the state at that moment.

What do you receive when the assessment ends?

The product is a report with separate parts for separate readers: an executive summary, technical detail for developers, and a ranked list of fixes. Owners see how things stand overall and what to fix first. Developers read which door fails, what happens and how to fix it.

Once the fixes are in, one follow-up assessment costs you nothing. It checks that the places you thought were fixed are closed, and the result goes into the report. Fixing and re-checking are inside the price, so the work does not stop at diagnosis.

What does it cost, and what changes the amount?

The price is set mostly by the number of doors and the combinations of permission to be tested. All amounts exclude tax.

Item Price Range Included
API assessment from JPY 98,000 (excl. tax) up to 10 endpoints Tool scan and a person's review, a report, one free re-assessment after fixes
Individual vulnerability fix from JPY 10,000 per item (excl. tax) Holes found in the assessment Quoted by content
Free assessment JPY 0 After the NDA and the consent form A quick tool scan plus a person's review, with an issue report

An endpoint is one door counted on its own. "Return the order list" and "update member information" each count as one, one per operation. Beyond 10, we look at your list of doors and quote.

An API with many kinds of permission, or with many connections to outside services, has more combinations to test. We take that into account in the quote. A formal quotation follows the free assessment.

Every price is listed on the pricing page. Run the free URL check first for the outside view of your public website. If the number of doors or the authentication method is unclear, contact us and tell us your situation.

How does an order proceed from first contact to delivery?

This follows the general flow from first contact to after-delivery support. Below are the stages that matter for an API assessment.

  1. In the free consultation we ask about the number of endpoints, the authentication method and the kinds of permission.
  2. Before work starts we sign a non-disclosure agreement (NDA) at no charge.
  3. Once you have signed the consent form, we run the free assessment and give you the result as an issue report.
  4. The issues found are turned into a fix estimate. A full rebuild, if needed, gets its own estimate.
  5. After your order, we assess all items and close the holes.
  6. Ahead of delivery we share a PDF of what was fixed and how, and you confirm the behavior yourself.
  7. We deliver. You get one month of free support afterward.

If you want to keep watching after the assessment, the maintenance and advisory plans and our AI monitoring service, Guard, are available.

Our own developers test the API. Is an outside assessment still needed?

Developer tests matter, and an assessment builds on them. Your team knows how each endpoint is meant to behave, and that is where the blind spot lies. An unexpected combination of permissions, or an old door nobody uses any more, stands out to a reviewer who did not write it.

For your own review, the OWASP API Security Top 10 works as a checklist. Bring in an outside reviewer before a release or after a large specification change.

Frequently asked questions

What is an API assessment?

It is a check from the outside on the doors (APIs) through which your systems trade data with smartphone apps and partner services, looking for gaps in permission checks, authentication and usage limits. Tools scan broadly. Then a person checks the permission boundaries by hand, and the results go into a report.

How much does it cost?

It is from JPY 98,000 (excl. tax, up to 10 endpoints). Beyond that range, we look at your list of endpoints and quote. An individual vulnerability fix is from JPY 10,000 per item (excl. tax), quoted by content, and one re-assessment after the fixes is free.

We already had a web application assessment. Do we still need an API assessment?

Even when a web screen and a smartphone app share the same API, what the screen shows and what the API returns are not always the same. A screen assessment looks for holes in the screen, and an API assessment looks for holes in the door itself, so their roles differ.

Could the assessment corrupt live data or stop the service?

The consent form lists the target endpoints, the period and the methods in writing ahead of the work. For operations that worry you on a live system, such as payment or deleting data, we agree how to handle them beforehand. If you have a test environment, we can also work there.

Does an assessment prove the API is safe?

No. An assessment tells you which holes were found at that moment. New holes can appear when endpoints are added or specifications change, so it matters to check again after each change.

Next step

If you first want to see how your public site appears from outside, use the free URL check. Type in a URL and it looks only at what any visitor could see. It does not run attacks, and it cannot see how your API checks permissions.

If the number of doors, the authentication method or partner connections are unclear, contact us by form, phone or online meeting. We sign the NDA, at no cost, before the assessment starts.

Written by Mio Kanzaki (AI security analyst). Supervised by 千葉圭介. Updated 19 November 2026.

Frequently asked questions

What is an API assessment?

It is a check from the outside on the doors (APIs) through which your systems trade data with smartphone apps and partner services, looking for gaps in permission checks, authentication and usage limits. Tools scan broadly. Then a person checks the permission boundaries by hand, and the results go into a report.

How much does it cost?

It is from JPY 98,000 (excl. tax, up to 10 endpoints). Beyond that range, we look at your list of endpoints and quote. An individual vulnerability fix is from JPY 10,000 per item (excl. tax), quoted by content, and one re-assessment after the fixes is free.

We already had a web application assessment. Do we still need an API assessment?

Even when a web screen and a smartphone app share the same API, what the screen shows and what the API returns are not always the same. A screen assessment looks for holes in the screen, and an API assessment looks for holes in the door itself, so their roles differ.

Could the assessment corrupt live data or stop the service?

The consent form lists the target endpoints, the period and the methods in writing ahead of the work. For operations that worry you on a live system, such as payment or deleting data, we agree how to handle them beforehand. If you have a test environment, we can also work there.

Does an assessment prove the API is safe?

No. An assessment tells you which holes were found at that moment. New holes can appear when endpoints are added or specifications change, so it matters to check again after each change.

See what an attacker can see, today

The free URL check returns a grade in under a minute. For costs and next steps, book a free consultation.