Skip to main content
Comparison · Aikido

B-52 and Aikido meet at the AI pentest

Aikido is a platform for code, cloud and runtime security that also sells AI pentesting. B-52 is a penetration testing platform. What follows covers where the two pentests overlap, which classes each one publishes, how a report gets signed, and what each side puts a price on. Every statement about Aikido was read on aikido.dev on 18 November 2026.

Their published position was last read on 2026-11-18. Every claim below carries the page it came from and the date it was checked.

What aikido sells

A code, cloud and runtime security platform, with AI pentesting inside it

Their account of their own product comes first, in their terms. Where a phrase is theirs it is quoted in the claims table at the foot of this page.

Aikido describes itself as one system for securing code, cloud and runtime. Its homepage sets out four product areas. The first covers code: dependency scanning, static analysis, secrets detection and code quality, with automated fixes. The second covers cloud: visibility into live workloads, services, misconfigurations and data flows. The third is attack, where autonomous agents attack running applications, APIs and infrastructure to prove what is exploitable. The fourth is protection against supply chain attacks on developer devices. AI pentesting sits in the third area, and it is sold in two ways: as a pentest bought by itself, with a report at the end, or through the monthly credits a paid platform plan includes. So a buyer comparing the two products is comparing one part of Aikido with the whole of B-52. B-52 is a penetration testing platform from Security Brigade: eleven coverage classes, three delivery models, and a report signable under Security Brigade’s CERT-In empanelment in every one of them. The rest of this page keeps to the ground the two share, and names the parts of each product that fall outside it.

In their own terms

What Aikido publishes, part by part

Their pages, their names. Each row is sourced in the claims table, with the page it was read on and the date.

PartWhat it is, as published
Code Dependency scanning, static analysis, secrets detection and code quality checks, with automated fixes. A Code Audit product traces data flow, ownership checks and permission boundaries across a repository.
Cloud Visibility into live workloads, services, misconfigurations and data flows, published as posture management and scanning.
AI Pentesting Autonomous agents that discover and exploit flaws across applications, APIs and infrastructure. Four targets are named: web applications, network, mobile applications and AI agents.
This is the part the rest of the page compares.
Aikido Infinite Continuous pentesting: a push to staging triggers a scoped pentest of the changed code and the surfaces it affects, and a confirmed finding comes with a pull request carrying the fix.
Aikido Machine A GPU server installed on the customer’s premises that runs continuous penetration tests on applications, with no internet connection required, sold for an annual fee.
Human-led pentest Offered for the standards and auditors that require a test performed by a person, with AI pentests on each release in between.
Plans Free, Basic, Pro and Advanced, with tailored pricing for enterprises. Paid plans include monthly credits that are spent on AI features, AI pentesting among them.

Code

What it is, as published
Dependency scanning, static analysis, secrets detection and code quality checks, with automated fixes. A Code Audit product traces data flow, ownership checks and permission boundaries across a repository.

Cloud

What it is, as published
Visibility into live workloads, services, misconfigurations and data flows, published as posture management and scanning.

AI Pentesting

What it is, as published
Autonomous agents that discover and exploit flaws across applications, APIs and infrastructure. Four targets are named: web applications, network, mobile applications and AI agents.

This is the part the rest of the page compares.

Aikido Infinite

What it is, as published
Continuous pentesting: a push to staging triggers a scoped pentest of the changed code and the surfaces it affects, and a confirmed finding comes with a pull request carrying the fix.

Aikido Machine

What it is, as published
A GPU server installed on the customer’s premises that runs continuous penetration tests on applications, with no internet connection required, sold for an annual fee.

Human-led pentest

What it is, as published
Offered for the standards and auditors that require a test performed by a person, with AI pentests on each release in between.

Plans

What it is, as published
Free, Basic, Pro and Advanced, with tailored pricing for enterprises. Paid plans include monthly credits that are spent on AI features, AI pentesting among them.

Coverage

Eleven classes, read against what Aikido publishes as a pentest

Their column is what their pages name as a pentest target today. B-52 covers all eleven in all three of its delivery models.

Coverage classOn Aikido, as publishedOn B-52
Web application Published. The standard pentest covers front-end applications, web applications and their APIs. Covered, including the authenticated flows and the authorisation decisions behind each role, with a credential supplied per role.
API Published, in the same pentest: REST, GraphQL, gRPC and SOAP are named. Covered.
Mobile application Published for Android and iOS. Their page states that the mobile source code is required, and that an Android build cannot use certificate pinning, root detection, emulator detection or runtime self-protection. Covered. Release builds are tested as published: obfuscated, certificate-pinned and defended at runtime, with no source required.
Their page describes mobile pentests as whitebox only for now. Read 2026-11-18.
External network Published. Starting from one domain, agents find each internet-facing system and pentest it, then test again when something changes. Covered, with the perimeter worked out from announced ranges, resolving names and what answers on each address, then proposed back to you as the scope.
LLM applications Published as a focused assessment of MCPs, chats, RAG, tools, memory and other LLM features. Covered on the platform.
Secure code review Published as products of the code area: static analysis in every plan, and a Code Audit run by agents across the repository. Covered. A file, a line, the path an attacker takes to reach that line, and the closing change in the repository’s own idiom.
Cloud Published as posture management and scanning of cloud accounts. A cloud penetration test is not named as a class on the pages read. Covered: the identity and permission layer, proved by making the call.
Internal network Published for applications that run on internal networks, reached through a broker or tested by a server on the customer’s premises. An internal network penetration test is not named as a class on the pages read. Covered. This is the one class that needs something deployed inside your network, and on-premise and customer-VPC deployments are both available for it.
Active Directory Not named on the pages read. Covered: privilege paths walked end to end.
Thick client Not named on the pages read. Covered on the platform.
Social engineering Not named on the pages read. Covered, and scoped separately from the other classes.
Dependency, secrets and cloud posture scanning Published in every plan, the free one included. Not offered.

Web application

On Aikido, as published
Published. The standard pentest covers front-end applications, web applications and their APIs.
On B-52
Covered, including the authenticated flows and the authorisation decisions behind each role, with a credential supplied per role.

API

On Aikido, as published
Published, in the same pentest: REST, GraphQL, gRPC and SOAP are named.
On B-52
Covered.

Mobile application

On Aikido, as published
Published for Android and iOS. Their page states that the mobile source code is required, and that an Android build cannot use certificate pinning, root detection, emulator detection or runtime self-protection.
On B-52
Covered. Release builds are tested as published: obfuscated, certificate-pinned and defended at runtime, with no source required.

Their page describes mobile pentests as whitebox only for now. Read 2026-11-18.

External network

On Aikido, as published
Published. Starting from one domain, agents find each internet-facing system and pentest it, then test again when something changes.
On B-52
Covered, with the perimeter worked out from announced ranges, resolving names and what answers on each address, then proposed back to you as the scope.

LLM applications

On Aikido, as published
Published as a focused assessment of MCPs, chats, RAG, tools, memory and other LLM features.
On B-52
Covered on the platform.

Secure code review

On Aikido, as published
Published as products of the code area: static analysis in every plan, and a Code Audit run by agents across the repository.
On B-52
Covered. A file, a line, the path an attacker takes to reach that line, and the closing change in the repository’s own idiom.

Cloud

On Aikido, as published
Published as posture management and scanning of cloud accounts. A cloud penetration test is not named as a class on the pages read.
On B-52
Covered: the identity and permission layer, proved by making the call.

Internal network

On Aikido, as published
Published for applications that run on internal networks, reached through a broker or tested by a server on the customer’s premises. An internal network penetration test is not named as a class on the pages read.
On B-52
Covered. This is the one class that needs something deployed inside your network, and on-premise and customer-VPC deployments are both available for it.

Active Directory

On Aikido, as published
Not named on the pages read.
On B-52
Covered: privilege paths walked end to end.

Thick client

On Aikido, as published
Not named on the pages read.
On B-52
Covered on the platform.

Social engineering

On Aikido, as published
Not named on the pages read.
On B-52
Covered, and scoped separately from the other classes.

Dependency, secrets and cloud posture scanning

On Aikido, as published
Published in every plan, the free one included.
On B-52
Not offered.

Three ways to run it

On demand, on every release, or inside your own walls

The same two products, read at the three points where a buyer chooses how a test is run.

01 One test

A single application, tested now

On Aikido
A standard pentest of one application and its primary APIs, with a PDF report, or a pentest scoped automatically from the connected repositories. Both are priced on their pricing page, and the figures are in the claims table below.
On B-52
B-52 publishes configurable pricing for web, API, mobile and cloud assessments. Other testing classes and delivery models are quoted individually.
What it decides
Which classes the single test has to reach, and who you want reviewing the findings before the report reaches you.
02 Every release

A test wired to the pipeline

On Aikido
Aikido Infinite triggers a scoped pentest on a push to staging. Separately, their build checks run in GitHub, GitLab, Bitbucket, CircleCI and Jenkins and can block a merge on a critical issue.
On B-52
Four continuous-integration systems in production: GitHub Actions, GitLab CI, Jenkins and Azure DevOps. A result can fail a build against a severity threshold you set.
What it decides
Whether the gate you want is a scanner on each build, a pentest on each release, or both.
03 Inside the network

Code and findings that cannot leave

On Aikido
Aikido Machine: a GPU server installed in your data centre that runs continuous penetration tests on applications with no internet connection required, for an annual fee.
On B-52
On-premise and customer VPC are both available today, with residency in India, the European Union, the United States and Singapore. The models can run on standard hosting, on hardware dedicated to you or inside your own environment, or you can bring your own enterprise model.
What it decides
Where the work runs and where the data sits. Both products publish an answer for a buyer whose code cannot leave the building.

Whose signature

Who verifies a finding, and whose name the report goes out under

Aikido’s AI pentest produces a report its page describes as audit-ready, structured for the compliance frameworks named in the claims table. For the standards and auditors that require a test performed by a person, Aikido offers a human-led pentest as a separate engagement. B-52 has three delivery models over one coverage set, and the report from every one of them is signable under Security Brigade’s CERT-In empanelment. In the fully autonomous model you sign off the scope and the run produces the report, with the proof on every finding and an independent automated cross-check on top. In the autonomous, expert-verified model a senior Security Brigade auditor verifies every finding before it reaches you. In the human-led model a Security Brigade team works the engagement with the platform underneath it. Security Brigade has been CERT-In empanelled since 2008 and is ISO 27001 certified. The eleven classes are the same in all three; what changes is who reviews the work.

What each side can show

Six questions a buyer asks at this stage, answered for both

Their column is sourced in the claims table below. Ours is what this site publishes everywhere else.

The questionAikidoB-52
What arrives with a finding Impact, reproduction steps and remediation guidance, with an automated fix where one is possible. Findings are reported only after they are exploited and confirmed against the live target. The request as sent and the response as returned, the steps that reproduce it, a CVSS v4.0 vector and a CWE, on every finding, in every class, in all three delivery models.
What the test needs from you Code access is optional for a web pentest and, in their words, improves results. Black-box and grey-box tests are available at extra cost. A mobile pentest requires the source code. A credential for each role you want exercised. Nothing is required inside your network for external and application testing, and a deployment only for internal.
A measured comparison against human testers Their FAQ states that for web applications AI pentesting delivers coverage comparable to a traditional human-led pentest. No method is published beside the statement on that page. B-52 and Security Brigade’s expert assessment team worked the same targets in parallel, findings pooled with each item counted once. B-52 reached 90–95% of that pooled set. The method is set out below.
A public benchmark Not evaluated in this comparison. B-52 completed all 104 XBOW validation challenges with no human steering after each target URL was supplied.
How long a result takes A report in a few hours, per their AI pentesting page. An observed median of one to three business days from scope sign-off to report, in the fully autonomous model.
Customer evidence Published: a count of organisations and developers on the homepage. Anonymised case studies drawn from real B-52 engagements, each attack chain reconstructed step by step.

What arrives with a finding

Aikido
Impact, reproduction steps and remediation guidance, with an automated fix where one is possible. Findings are reported only after they are exploited and confirmed against the live target.
B-52
The request as sent and the response as returned, the steps that reproduce it, a CVSS v4.0 vector and a CWE, on every finding, in every class, in all three delivery models.

What the test needs from you

Aikido
Code access is optional for a web pentest and, in their words, improves results. Black-box and grey-box tests are available at extra cost. A mobile pentest requires the source code.
B-52
A credential for each role you want exercised. Nothing is required inside your network for external and application testing, and a deployment only for internal.

A measured comparison against human testers

Aikido
Their FAQ states that for web applications AI pentesting delivers coverage comparable to a traditional human-led pentest. No method is published beside the statement on that page.
B-52
B-52 and Security Brigade’s expert assessment team worked the same targets in parallel, findings pooled with each item counted once. B-52 reached 90–95% of that pooled set. The method is set out below.

A public benchmark

Aikido
Not evaluated in this comparison.
B-52
B-52 completed all 104 XBOW validation challenges with no human steering after each target URL was supplied.

How long a result takes

Aikido
A report in a few hours, per their AI pentesting page.
B-52
An observed median of one to three business days from scope sign-off to report, in the fully autonomous model.

Customer evidence

Aikido
Published: a count of organisations and developers on the homepage.
B-52
Anonymised case studies drawn from real B-52 engagements, each attack chain reconstructed step by step.

How the parallel benchmark was run

The 90–95% figure, and the set it is a share of

Measured against our own assessors As of 2026-09-14
  • B-52 and Security Brigade’s expert assessment team worked the same targets at the same time, and neither side saw the other’s output while the work was running.
  • Both sets of findings were pooled into one denominator with each item counted once, so a defect both sides reached counts once, not twice.
  • B-52 reached 90–95% of that pooled set. Part of what it reached was absent from the team’s own output, which is why the pooled set is larger than either side produced alone.
  • Every engagement Security Brigade has run since the firm started in 2006 was worked inside Lemon, and that record is what trained the models the platform runs on.

FAQ

Common questions

Still deciding whether this is the right test? Ask us directly.

Request a scoping call
What is Aikido Security?+
A platform for code, cloud and runtime security. Its homepage describes four product areas: code scanning with automated fixes, cloud visibility, autonomous attack testing, and protection against supply chain attacks on developer devices. AI pentesting is one product inside the attack area, sold by itself or through the credits a paid plan includes.
Does Aikido do penetration testing?+
Yes. Aikido sells AI pentesting for web applications and APIs, external networks, mobile applications and LLM features, a continuous version called Aikido Infinite, and an on-premise server called Aikido Machine. It also offers a human-led pentest for the standards and auditors that require one. Each of those is sourced and dated in the claims table on this page.
How much does an Aikido pentest cost?+
Their pricing page publishes a figure for a standard pentest of one application and a range for a pentest scoped from your repositories. Both figures, the page they were read on and the date are in the claims table on this page. B-52 publishes configurable pricing for web, API, mobile and cloud assessments. Other testing classes and delivery models are quoted individually.
What does B-52 cover that Aikido does not publish as a pentest?+
Thick-client applications, Active Directory and social engineering are not named on the Aikido pages read for this comparison. Cloud is published by Aikido as posture management and scanning, and internal testing is published for applications on internal networks; B-52 covers cloud and internal network as penetration testing classes. Aikido also publishes dependency, secrets and cloud posture scanning.
Can a B-52 report be filed where an empanelled auditor is required?+
Yes, from all three delivery models. Every B-52 report is signable under Security Brigade’s CERT-In empanelment and usable as audit evidence, and Security Brigade has been CERT-In empanelled since 2008. The models differ in who reviews the work: an independent automated cross-check, a senior auditor verifying every finding, or a senior auditor running the engagement.
How current is this comparison?+
Every claim on this page was read on Aikido’s own pages on 18 November 2026 and carries that date beside it in the table at the foot of the page. The row most likely to move first is mobile: their page describes mobile pentests as whitebox only for now.

Every claim, sourced

What was read, and when

Competitive claims go out of date, and this page has a date on every one of them so you can tell how far. Where a row has aged past what you would rely on, check it against their own site. That is where each of these was read.

What Aikido publishesVerified
Homepage positioning: "Continuous, autonomous security that gets developers back to building." and "Secure your code, cloud, and runtime in one central system." The page states "Join over 50k orgs and 100K devs worldwide". 2026-11-18
Read at https://www.aikido.dev/
Four product areas are described on the homepage. Code: "Discover and AutoFix vulnerabilities, malicious dependencies, secrets and code quality issues before they ship." Cloud: "Gain real-time visibility into live workloads, services, misconfigurations, and data flows across your environments." Attack: "Autonomously attack your running applications, APIs and infrastructure to prove what’s actually exploitable, before others do." Protection: "Block supply chain attacks before they hit your devices." 2026-11-18
Read at https://www.aikido.dev/
The platform is sold in plans named Free, Basic, Pro and Advanced, with "tailored pricing" for enterprises. Free is "free forever" and includes 2 users. Features listed in Free include "Dependency Scanning (SCA)", "SAST & AI SAST", "Secrets Detection" and "DAST". Paid plans include monthly credits: "Credits are used for advanced AI features such as AI Pentesting, Deep PR Reviews, and AI Code Analysis." 2026-11-18
Read at https://www.aikido.dev/pricing
AI Pentesting, in their words: "Hundreds of autonomous agents act like real pentesters, discovering and exploiting flaws across your apps, APIs and infrastructure." The page names four targets: web apps, network, mobile apps and AI agents. The standard pentest "Pentests frontend apps, web apps, and APIs (REST, GraphQL, gRPC, SOAP,...)". 2026-11-18
Read at https://www.aikido.dev/attack/aipentest
Proof per finding: "Unproven issues are dropped. Verified findings ship with impact, reproduction steps, remediation guidance and AutoFixes where possible." and "Findings are only reported after they are successfully exploited and confirmed against the live target." 2026-11-18
Read at https://www.aikido.dev/attack/aipentest
Source code: asked whether code access is needed, the page answers "No, but providing code access significantly improves results." The standard pentest is listed with "Black-box and grey-box testing also available (extra cost due to more agents required)". 2026-11-18
Read at https://www.aikido.dev/attack/aipentest
Scope and safety controls: "You choose which domains can be attacked. Pre-flight checks before every run." "Every step, action or request is logged and visible to you." "When Aikido finds a vulnerability that could be escalated, it pauses and shows you the full attack analysis before it goes further." 2026-11-18
Read at https://www.aikido.dev/attack/aipentest
Pentest pricing as published: a "TYPICAL Pentest" at $4,000 "per assessment", described as a "Time-boxed, fixed-scope security audit for a single application and its primary APIs", and a "Rightsized Pentest" at "$50 - $30,000+", "scoped to your application". The page states "No High or Critical Finding = Don’t Pay" and offers re-testing "for up to 6 months". 2026-11-18
Read at https://www.aikido.dev/pricing
Compliance use of the report: "Get a full audit-grade SOC2 or ISO27001 PDF report in a few hours." "Every run produces an audit-ready penetration test report with validated findings, proof-of-exploit details, and remediation guidance, structured to meet SOC 2 and ISO 27001 requirements." Report types named: a full auditor report "for SOC 2, ISO 27001 and HIPAA", a management report, a customer-facing report and a post-remediation report. 2026-11-18
Read at https://www.aikido.dev/attack/aipentest
Human-led testing: "Some standards and auditors still require a pentest performed by a human. Aikido is CREST accredited for penetration testing, so you can get a human-led pentest from us when you need one, and run AI pentests on every release in between." 2026-11-18
Read at https://www.aikido.dev/attack/aipentest
Comparison with human testers, from their FAQ: "For web applications, AI Pentesting delivers coverage comparable to a traditional human-led pentest, with results available in hours instead of weeks." and "Human testers remain valuable for non-web targets and highly contextual edge cases." 2026-11-18
Read at https://www.aikido.dev/attack/aipentest
Continuous pentesting is sold as Aikido Infinite: "Every push to staging triggers a scoped pentest, targeting changed code and affected surfaces automatically." Confirmed vulnerabilities get "a merge-ready PR with the code-level fix". 2026-11-18
Read at https://www.aikido.dev/attack/infinite
Mobile pentesting: "Run Android and iOS in the same pentest, along with the backend they share". "The mobile source code is required, since mobile pentests are whitebox only for now, on both Android and iOS." "For Android, the build can’t use certificate pinning, and it can’t have root detection, emulator detection or any runtime application self-protection (RASP) enabled." "For iOS, upload a signed IPA that isn’t encrypted." 2026-11-18
Read at https://www.aikido.dev/attack/mobile-pentesting
External network pentesting: "Starting from just one domain, our AI agents find every internet-facing system you run and pentest each one, then test again whenever something changes." "Agents pull leaked credentials from the dark web and try them on every login they find." 2026-11-18
Read at https://www.aikido.dev/attack/external-network-pentesting
AI and LLM features as a pentest target: "Get a focused assessment of MCPs, chats, RAG, tools, memory, or other LLM features." The standard pentest lists "Prompt Injection" among what it tests for. 2026-11-18
Read at https://www.aikido.dev/attack/aipentest
On-premise: "Aikido Machine is a GPU server that runs continuous penetration tests on your critical applications, fully within your premises. For regulated industries." "Aikido Machine runs inside your network with no internet connection required." "Aikido charges one annual fee for the software, upgrades, support, and the service level agreement (SLA)." The Advanced plan lists "Broker for Internal Apps". 2026-11-18
Read at https://www.aikido.dev/attack/aikido-machine
Pipelines: "Wire Aikido into your existing CI/CD and it scans every build for vulnerabilities, leaked secrets, and supply chain risks. Critical issues block the merge, everything else ships." Systems named on the page: GitHub, GitLab, Bitbucket, CircleCI and Jenkins. 2026-11-18
Read at https://www.aikido.dev/code/ci-cd-pipeline-security
Code analysis: static analysis is in every plan, and a Code Audit product is described as "The audit traces data flow, ownership checks, permission boundaries, and service interactions to catch where the logic breaks down, not just where a single line looks off." "Each finding shows what’s vulnerable, why it’s exploitable, and how an attacker would reach it, with a full reasoning trace." 2026-11-18
Read at https://www.aikido.dev/code/code-audit
Their own certifications, as their trust page states them: "Aikido is ISO 27001:2022, ISO 42001 & AICPA’s SOC 2 Type II compliant, and FedRAMP Certified: Class C (Moderate)." and "Aikido is CREST accredited for penetration testing. Our human-led pentests follow processes CREST has independently reviewed." 2026-11-18
Read at https://www.aikido.dev/trust-center
Thick-client applications, Active Directory and social engineering are not named on the homepage, the AI Pentesting page, the mobile, external network and Aikido Machine pages, or the pricing page. Internal testing is published for applications: the pricing page describes its broker as "Pentest apps running on internal networks". 2026-11-18
Read at https://www.aikido.dev/pricing

Homepage positioning: "Continuous, autonomous security that gets developers back to building." and "Secure your code, cloud, and runtime in one central system." The page states "Join over 50k orgs and 100K devs worldwide".

Verified
2026-11-18

Read at https://www.aikido.dev/

Four product areas are described on the homepage. Code: "Discover and AutoFix vulnerabilities, malicious dependencies, secrets and code quality issues before they ship." Cloud: "Gain real-time visibility into live workloads, services, misconfigurations, and data flows across your environments." Attack: "Autonomously attack your running applications, APIs and infrastructure to prove what’s actually exploitable, before others do." Protection: "Block supply chain attacks before they hit your devices."

Verified
2026-11-18

Read at https://www.aikido.dev/

The platform is sold in plans named Free, Basic, Pro and Advanced, with "tailored pricing" for enterprises. Free is "free forever" and includes 2 users. Features listed in Free include "Dependency Scanning (SCA)", "SAST & AI SAST", "Secrets Detection" and "DAST". Paid plans include monthly credits: "Credits are used for advanced AI features such as AI Pentesting, Deep PR Reviews, and AI Code Analysis."

Verified
2026-11-18

Read at https://www.aikido.dev/pricing

AI Pentesting, in their words: "Hundreds of autonomous agents act like real pentesters, discovering and exploiting flaws across your apps, APIs and infrastructure." The page names four targets: web apps, network, mobile apps and AI agents. The standard pentest "Pentests frontend apps, web apps, and APIs (REST, GraphQL, gRPC, SOAP,...)".

Verified
2026-11-18

Read at https://www.aikido.dev/attack/aipentest

Proof per finding: "Unproven issues are dropped. Verified findings ship with impact, reproduction steps, remediation guidance and AutoFixes where possible." and "Findings are only reported after they are successfully exploited and confirmed against the live target."

Verified
2026-11-18

Read at https://www.aikido.dev/attack/aipentest

Source code: asked whether code access is needed, the page answers "No, but providing code access significantly improves results." The standard pentest is listed with "Black-box and grey-box testing also available (extra cost due to more agents required)".

Verified
2026-11-18

Read at https://www.aikido.dev/attack/aipentest

Scope and safety controls: "You choose which domains can be attacked. Pre-flight checks before every run." "Every step, action or request is logged and visible to you." "When Aikido finds a vulnerability that could be escalated, it pauses and shows you the full attack analysis before it goes further."

Verified
2026-11-18

Read at https://www.aikido.dev/attack/aipentest

Pentest pricing as published: a "TYPICAL Pentest" at $4,000 "per assessment", described as a "Time-boxed, fixed-scope security audit for a single application and its primary APIs", and a "Rightsized Pentest" at "$50 - $30,000+", "scoped to your application". The page states "No High or Critical Finding = Don’t Pay" and offers re-testing "for up to 6 months".

Verified
2026-11-18

Read at https://www.aikido.dev/pricing

Compliance use of the report: "Get a full audit-grade SOC2 or ISO27001 PDF report in a few hours." "Every run produces an audit-ready penetration test report with validated findings, proof-of-exploit details, and remediation guidance, structured to meet SOC 2 and ISO 27001 requirements." Report types named: a full auditor report "for SOC 2, ISO 27001 and HIPAA", a management report, a customer-facing report and a post-remediation report.

Verified
2026-11-18

Read at https://www.aikido.dev/attack/aipentest

Human-led testing: "Some standards and auditors still require a pentest performed by a human. Aikido is CREST accredited for penetration testing, so you can get a human-led pentest from us when you need one, and run AI pentests on every release in between."

Verified
2026-11-18

Read at https://www.aikido.dev/attack/aipentest

Comparison with human testers, from their FAQ: "For web applications, AI Pentesting delivers coverage comparable to a traditional human-led pentest, with results available in hours instead of weeks." and "Human testers remain valuable for non-web targets and highly contextual edge cases."

Verified
2026-11-18

Read at https://www.aikido.dev/attack/aipentest

Continuous pentesting is sold as Aikido Infinite: "Every push to staging triggers a scoped pentest, targeting changed code and affected surfaces automatically." Confirmed vulnerabilities get "a merge-ready PR with the code-level fix".

Verified
2026-11-18

Read at https://www.aikido.dev/attack/infinite

Mobile pentesting: "Run Android and iOS in the same pentest, along with the backend they share". "The mobile source code is required, since mobile pentests are whitebox only for now, on both Android and iOS." "For Android, the build can’t use certificate pinning, and it can’t have root detection, emulator detection or any runtime application self-protection (RASP) enabled." "For iOS, upload a signed IPA that isn’t encrypted."

Verified
2026-11-18

Read at https://www.aikido.dev/attack/mobile-pentesting

External network pentesting: "Starting from just one domain, our AI agents find every internet-facing system you run and pentest each one, then test again whenever something changes." "Agents pull leaked credentials from the dark web and try them on every login they find."

Verified
2026-11-18

Read at https://www.aikido.dev/attack/external-network-pentesting

AI and LLM features as a pentest target: "Get a focused assessment of MCPs, chats, RAG, tools, memory, or other LLM features." The standard pentest lists "Prompt Injection" among what it tests for.

Verified
2026-11-18

Read at https://www.aikido.dev/attack/aipentest

On-premise: "Aikido Machine is a GPU server that runs continuous penetration tests on your critical applications, fully within your premises. For regulated industries." "Aikido Machine runs inside your network with no internet connection required." "Aikido charges one annual fee for the software, upgrades, support, and the service level agreement (SLA)." The Advanced plan lists "Broker for Internal Apps".

Verified
2026-11-18

Read at https://www.aikido.dev/attack/aikido-machine

Pipelines: "Wire Aikido into your existing CI/CD and it scans every build for vulnerabilities, leaked secrets, and supply chain risks. Critical issues block the merge, everything else ships." Systems named on the page: GitHub, GitLab, Bitbucket, CircleCI and Jenkins.

Verified
2026-11-18

Read at https://www.aikido.dev/code/ci-cd-pipeline-security

Code analysis: static analysis is in every plan, and a Code Audit product is described as "The audit traces data flow, ownership checks, permission boundaries, and service interactions to catch where the logic breaks down, not just where a single line looks off." "Each finding shows what’s vulnerable, why it’s exploitable, and how an attacker would reach it, with a full reasoning trace."

Verified
2026-11-18

Read at https://www.aikido.dev/code/code-audit

Their own certifications, as their trust page states them: "Aikido is ISO 27001:2022, ISO 42001 & AICPA’s SOC 2 Type II compliant, and FedRAMP Certified: Class C (Moderate)." and "Aikido is CREST accredited for penetration testing. Our human-led pentests follow processes CREST has independently reviewed."

Verified
2026-11-18

Read at https://www.aikido.dev/trust-center

Thick-client applications, Active Directory and social engineering are not named on the homepage, the AI Pentesting page, the mobile, external network and Aikido Machine pages, or the pricing page. Internal testing is published for applications: the pricing page describes its broker as "Pentest apps running on internal networks".

Verified
2026-11-18

Read at https://www.aikido.dev/pricing

Run one application, and read the report it returns

B-52 publishes configurable pricing for web, API, mobile and cloud assessments. Other testing classes and delivery models are quoted individually.