Commercial platform

SwarmAttacker vs Aikido Attack

Aikido Security's hosted, closed-source AI pentesting product that launches up to hundreds of sandboxed agents against a scoped app, preferably with the source connected.

Vendor
Aikido Security
Licence
Proprietary, SaaS only
Pricing
Credit-based, one credit is one dollar. A standard pentest is $4,000 and includes retests.
Checked
September 2, 2026
CapabilitySwarmAttacker8 of 8Aikido Attack1 of 8
FreeNot found.
Runs on a ChatGPT subscriptionNot applicable.
Custom prompts & skillsPartial.
Safety-refusal handlingNot applicable.
Parallel multi-agent swarmYes.
MIT-licensed open sourceNot found.
Black-box from a URLPartial.
Model-agnosticNot found.
  • yes
  • not found or only partly

Hover a mark for the evidence. A cross means the capability was not found, or only partly found, on Aikido Attack's public pages or repository on September 2, 2026, not that it is technically impossible.

Profile

What is Aikido Attack?

Aikido Attack is the AI pentesting product of Aikido Security, a Ghent-based application-security vendor valued at a billion dollars after a $60 million Series B in January 2026. Attack was formed from two 2025 acquisitions and announced that November, and has since added MCP-server testing, continuous per-deploy testing and Android testing.

It is a hosted, closed-source service. A control plane orchestrates up to a couple of hundred specialised agents, recon, mapping, CVE, exploit, web-vulnerability and logic-probing agents, each in an isolated sandbox with a browser, terminal and HTTP client, with scope enforced at the network level. Agents stop once a finding is proven; deeper exploitation requires a user to click through. Aikido describes itself as white-box first: up to six repositories, API specs, docs and free-text notes can be attached, and its own docs say that without code the system scans blind. The models are undisclosed and vendor-run, with no bring-your-own option.

Pricing is credit-based at a dollar per credit: a standard pentest is $4,000, and continuous runs cost ten credits per launched agent. Pentests are not part of the free platform plan. Independent evidence comes mainly from an Aikido-sponsored Doyensec study in May 2026 that, with source and credentials supplied, reported 49 verified findings across two open-source apps, at $4,000 each in under nine hours. Doyensec also noted a third of one app's severities were overstated.

In SwarmAttacker's favour

Where it stands out.

  • 01Free and open source, where a standard Aikido pentest is $4,000 and Attack is not in the free plan.
  • 02Runs from your machine on a ChatGPT plan, with model choice; Aikido is hosted with undisclosed, vendor-run models.
  • 03Black-box from a URL, where Aikido is white-box first and its own docs say it scans blind without source.
  • 04User-injectable attack skills and refusal handling, where Aikido offers only scope and notes fields and publishes nothing on refusals.

In Aikido Attack's favour

Where it is stronger.

  • 01Strong real-app evidence: 49 verified findings with a 4% false-positive rate in the Doyensec study.
  • 02Authenticated, code-aware testing with multi-role credentials, MFA, MCP servers and Android builds.
  • 03A hosted control plane with per-agent sandboxes, network-level scope enforcement and pause controls.
  • 04Compliance-ready reports and auto-fix inside a full code-to-cloud AppSec platform.
  • 05A well-funded, fast-moving company at a billion-dollar valuation.

Verdict

Which one should you pick?

Pick Aikido Attack if you want a hosted, code-aware service with authenticated testing, compliance reports and a vendor behind it, and $4,000 per assessment is acceptable. Pick SwarmAttacker if you want a free, open-source black-box agent you run yourself on a ChatGPT plan.

Turn a URL into a security report.

Open source, MIT licensed, and it runs on the ChatGPT plan you already pay for. Install it and point it at a target you own.

brew install joloooo/swarm/swarm
Source on GitHub