What do Dreadnode and Mindgard sell?

Dreadnode sells AI Red Teaming and Dreadnode Platform. Mindgard sells AI Red Teaming and AI Security Expert.

Are Dreadnode and Mindgard in the same category?

Yes. AI model and agent security is the primary category of both on CISO Tribune.

Who are Dreadnode and Mindgard sold to?

Dreadnode sells to organizations that require scaled agentic workloads and custom capabilities. Mindgard addresses organizations that secure the AI systems they build, buy and use.

Vendor comparison · 2 companies

Dreadnode vs Mindgard

Dreadnode sells an infrastructure platform that gives security teams the means to own, operationalize, and improve agentic cyber capabilities. Mindgard secures AI models, agents, and applications against exploitable threats. CISO Tribune files both under AI model and agent security. This page sets the two records side by side: what each sells, where it is based, who owns and funds it, the customers it names and its recent news, each fact numbered to its source.

Change the companies

Up to four companies side by side.

Dreadnode
Mindgard

Add a company

Dreadnode vs Mindgard: Products, Customers and Funding Compared
What it sells
	DreadnodeDreadnode sells an infrastructure platform that gives security teams the means to own, operationalize, and improve agentic cyber capabilities.dreadnode.io	MindgardMindgard secures AI models, agents, and applications against exploitable threats.mindgard.ai
What it sells	AI Red Teaming^[2][3][5]AI red teaming across AI systems in production, from traditional machine learning models to multi-agent systems. Dreadnode Platform^[3][4]Infrastructure for security teams to own, operationalize, and improve agentic cyber capabilities.	AI Red Teaming^{[1][4][6][10]}Mindgard's agentic platform red teams AI systems, agents, and infrastructure. AI Security Expert^[9]A subscription that gives a team direct access to a specialist.
Categories
Categories	AI model and agent security (primary category) Breach simulation and automated pen-testing Penetration testing and red team services	AI model and agent security (primary category) AI firewalls and guardrails
Headquarters
Headquarters	Not stated on its pages	Boston, United States^[2]
Founded
Founded	Not stated on its pages	2022^[2][3]
Ownership and funding
Ownership and funding	Not stated on its pages	Privately held^[2][3]
Named customers
Named customers	Not stated on its pages	Not stated on its pages
Recent news
Recent news	Not stated on its pages	August 13, 2026Mindgard announced a $30M Series A funding round.^[11] March 17, 2026Mindgard announced an expansion of its AI security platform that adds a reconnaissance capability for AI models, agents and applications.^[12]
G2 and Gartner Peer Insights
G2 and Gartner Peer Insights	G2: Not stated on its pages Gartner Peer Insights: Not stated on its pages	G2: Not stated on its pages Gartner Peer Insights: Not stated on its pages
Who it fits
Who it fits	Dreadnode sells to organizations that require scaled agentic workloads and custom capabilities. Dreadnode describes a cyber eval engagement for a frontier model provider. Organizations that need to run inside their own boundary can deploy Dreadnode on a Kubernetes cluster or in air-gapped environments, and on-prem deployment is available with the Enterprise plan.	Mindgard addresses organizations that secure the AI systems they build, buy and use. The offering is aimed at pen-testers who scale AI-focused security testing efforts, and at developers who integrate ongoing testing for secure AI deployments. Customer types shown include global insurance brokerages, healthcare AI platforms, global financial institutions, semiconductor manufacturers and enterprise technology services. Reporting is directed to security, engineering, GRC, and audit teams.
Last verified
Last verified	November 6, 2026	September 27, 2026

Where they overlap

AI model and agent security is the primary category of both Dreadnode and Mindgard.

Where they differ, field by field

Dreadnode is also filed under Breach simulation and automated pen-testing and Penetration testing and red team services. Mindgard is also filed under AI firewalls and guardrails.
Dreadnode lists AI Red Teaming and Dreadnode Platform. Mindgard lists AI Red Teaming and AI Security Expert.
Mindgard is headquartered in Boston, United States.
Mindgard was founded in 2022.
Mindgard is privately held.

How they differ

CISO Tribune files both Dreadnode and Mindgard under AI model and agent security as their primary category. Dreadnode also appears under breach simulation and automated pen-testing and under penetration testing and red team services, while Mindgard is also filed under AI firewalls and guardrails. Both sell AI red teaming, which is where their offerings overlap.

Dreadnode sells the Dreadnode Platform, infrastructure for security teams to own, operationalize, and improve agentic cyber capabilities. It supports workflows that coordinate agents, evaluations against published security tasks, and runtime policies that restrict tool access. Dreadnode can run on an existing Kubernetes cluster or in air-gapped environments, and offers on-prem deployment with the Enterprise plan. The one customer example it describes is a cyber eval engagement for a tier 1 frontier model provider. Mindgard sells an agentic platform that red teams AI systems, agents, and infrastructure, plus an AI Security Expert subscription that gives a team direct access to a specialist. Mindgard also tests whether AI guardrails and firewalls work.

Mindgard is headquartered in Boston, United States, was founded in 2022, and is privately held. It was created from AI security research at Lancaster University. On August 13, 2026, Mindgard announced a $30M Series A funding round. On March 17, 2026, it announced a platform expansion that adds a reconnaissance capability for AI models, agents and applications. Mindgard names customer types including global insurance brokerages, healthcare AI platforms, global financial institutions, and semiconductor manufacturers.

What it sells: Dreadnode

What AI Red Teaming does^[2][3][5]: Dreadnode applies AI red teaming across the full range of AI systems in production today, from traditional machine learning models to multi-agent systems. Dreadnode Platform evaluations probe models, applications, and deployed agents with adversarial attacks across supported languages and modalities.
What Dreadnode Platform does^[3][4]: The Dreadnode Platform equips defenders with ready-to-run capabilities and shared operational knowledge. Dreadnode can be deployed on an existing Kubernetes cluster or an embedded cluster on a dedicated VM, with offline installation bundles for air-gapped deployments.

What it sells: Mindgard

What AI Security Expert does^[9]: The expert shows how to run posture assessments, enhance tooling usage, build adversarial skills, and co-pilot red teaming efforts. The subscription includes workshops that show security and data science teams how adversarial techniques work, how agentic systems fail, and how defensive controls can be applied.

Questions about Dreadnode and Mindgard

What do Dreadnode and Mindgard sell?: Dreadnode sells AI Red Teaming and Dreadnode Platform. Mindgard sells AI Red Teaming and AI Security Expert.
Are Dreadnode and Mindgard in the same category?: Yes. AI model and agent security is the primary category of both on CISO Tribune.
Where are Dreadnode and Mindgard headquartered?: Mindgard is headquartered in Boston, United States.
When were Dreadnode and Mindgard founded?: Mindgard was founded in 2022.
Who owns Dreadnode and Mindgard?: Mindgard is privately held.
Who are Dreadnode and Mindgard sold to?: Dreadnode sells to organizations that require scaled agentic workloads and custom capabilities. Mindgard addresses organizations that secure the AI systems they build, buy and use.
What have Dreadnode and Mindgard announced recently?: August 13, 2026: Mindgard announced a $30M Series A funding round.

About Dreadnode

Dreadnode sells an infrastructure platform that gives security teams the means to own, operationalize, and improve agentic cyber capabilities. The Dreadnode Platform equips defenders with ready-to-run capabilities, shared operational knowledge, and the ability to run agents. Dreadnode describes the path to operationalize agents as fragmented and opaque. Dreadnode's capabilities are operator-authored agents for security testing, research, and operations, connected to the tools a team already uses.^[3][5]

Ready-to-run cyber capabilities on the Dreadnode Platform put agents to work on security testing, research, and operations. The Dreadnode Platform supports persistent sessions across web and terminal, with search of past work and sharing with a team. Dreadnode Platform workflows coordinate agents by sequencing tasks, running work in parallel, and branching or combining results. The Dreadnode Platform connects existing command-line tools, MCP servers, and custom integrations. Versioned capabilities on the Dreadnode Platform can be shared with their instructions, tools, and dependency setup.^[4]

Dreadnode Platform agents carry saved project context into new sessions and retrieve findings from earlier work. Dreadnode Platform users can review structured findings and assets with supporting evidence and relationships. The Dreadnode Platform versions and reuses datasets across evaluations, optimization, and training. Dreadnode Platform users can post-train models with demonstrations or rewards and retain the resulting model artifacts.^[4]

Dreadnode combines task evaluations, LLM judges, and AI red teaming so a team can set the criteria for greater responsibility. Dreadnode Platform evaluations benchmark agents against published security tasks and a team's own tasks and environments. Dreadnode uses LLM judges to assess correctness, scope adherence, and signs of cheating against a team's criteria. The Dreadnode Platform probes models, applications, and deployed agents with adversarial attacks across supported languages and modalities. The Dreadnode Platform compares models, prompts, and capability versions with per-sample results and judge reasoning.^[3][4]

Dreadnode Platform observability follows active agent sessions with live progress, metrics, and logs. The Dreadnode Platform inspects nested model and tool calls with their arguments and results. Dreadnode Platform traces connect findings back to the execution that produced them. Dreadnode Platform users can query traces to investigate failures and compare latency, token usage, and cost. Runtime policies in the Dreadnode Platform restrict tool access and check proposed actions against scope, and they allow, block, or request approval.^[4]

Dreadnode offers AI red teaming across AI systems in production today, from traditional machine learning models to multi-agent systems. Dreadnode offers web application penetration testing with 83 attack technique playbooks. Dreadnode describes an autonomous pentester with a built-in headless browser and 84 skills. Dreadnode agents run network discovery and Active Directory assessments, connected to attack-path analysis and C2 tools. Dreadnode's authored capabilities cover web, AI, network, reverse engineering, and forensics.^[2][3][4]

Dreadnode can be deployed on a customer's infrastructure, with application records, execution evidence, and capability artifacts in data stores the customer controls. Dreadnode runs on an existing Kubernetes cluster or on an embedded cluster on a dedicated VM, and offline installation bundles support air-gapped deployments. Dreadnode routes inference to approved providers or to models a customer hosts itself. Dreadnode offers on-prem deployment with the Enterprise plan. Dreadnode states that all evaluation, training, Worlds, and AI red teaming data stays within a customer's organization environment. Dreadnode SaaS meters workloads against a shared organization credit balance. Dreadnode stores secret values without returning them in normal API reads and injects them only into the remote runtimes and evaluations a user selects.^{[3][4][5][6][7]}

Dreadnode treats organizations as an ownership boundary, and workspaces as the main collaboration boundary inside an organization. In SaaS mode, creating or updating workspaces on Dreadnode requires a Pro plan or higher. Dreadnode is SOC 2, ISO 27001:2022, and ISO 27701 compliant. Dreadnode charges a one-time annual fee that varies based on deployment requirements, SLAs, and custom work. Dreadnode built an initial suite of ER-Spatial security tasks during a cyber eval engagement for a tier 1 frontier model provider.^{[3][4][5][8][9][10][11]}

Dreadnode on the record

About Mindgard

Mindgard secures AI models, agents, and applications against exploitable threats. Mindgard combines offensive security expertise, AI red teaming, attacker intelligence, and automated protection to help organizations discover, assess, and mitigate risk across AI models, agents, and applications. Mindgard was created from more than a decade of AI security research at Lancaster University. Mindgard was founded on pioneering research by Dr. Peter Garraghan at Lancaster University, which showed traditional AppSec could not address AI-specific risks. Mindgard's research has produced 150+ publicly disclosed AI vulnerabilities.^[1][5]

The Mindgard AI security platform finds, validates, and fixes high-impact AI vulnerabilities. Mindgard operates as an autonomous red teamer, continuously mapping, planning, and executing complex, agentic attack workflows. Mindgard profiles AI systems the way attackers do, mapping models, agents, tools, and behaviors before attack execution. Mindgard automates recon and intelligence gathering on AI systems before adversarial testing to discover instructions, tools, and behaviors attackers can exploit. Mindgard identifies AI models, agents, MCP/A2A servers, connected tools, and shadow AI within products. Mindgard tests AI applications, agents and LLMs, including image, audio and multi-modal models. Mindgard uses reconnaissance and agentic attacks to identify how an attacker could manipulate an agent, access sensitive information, misuse tools, bypass controls, or chain weaknesses together.^[1][4][13]

Mindgard validates which weaknesses are exploitable, explains their potential impact, and provides evidence and guidance to help teams close them. AI guardrails and firewalls apply policies to prompts, responses, or actions at runtime, and Mindgard tests whether those controls actually work. Mindgard assesses agents, tools, permissions, APIs, data, and infrastructure beyond the prompt layer. Mindgard evaluates and strengthens AI guardrails and WAF solutions against vulnerabilities.^[1][13]

Mindgard deploys through CI/CD, Burp Suite, or a single click. Mindgard integrates into existing workflows with CI/CD automation and a Burp Suite extension. Mindgard uses APIs and integrations to secure models, agents, and AI applications across development and production environments. Mindgard routes findings through APIs, CI/CD, and enterprise workflow integrations into existing security tooling, ticketing systems, and engineering workflows. Mindgard integrates report viewing into existing systems and SIEM. Mindgard supports SAML/SSO, SCIM provisioning, and RBAC to align security testing with organizational compliance standards.^{[1][4][13][14][15]}

Mindgard generates evidence, remediation guidance, and audit-ready reports mapped to OWASP, MITRE, NIST, and AIUC-1. Mindgard discovers and inventories AI applications, models, agents, and shadow AI across an entire environment. Mindgard prioritizes findings by severity, exposure, affected target, business context, and urgency of remediation. Mindgard gives security, engineering, GRC, and audit teams reporting they can use to demonstrate that AI risks and controls are being continuously evaluated. Security teams use Mindgard to discover and remediate high impact risk.^[4][15]

Mindgard researchers attack AI systems the way a motivated adversary would, using prompt injection, agent hijacking, tool abuse, data extraction and guardrail bypass, chained into attack paths. Mindgard red teams the whole AI system, not just the model. A single LLM application typically takes two to three weeks from kickoff to report in a Mindgard AI pentest. Mindgard sells a dedicated Technical Account Manager as an annual subscription. Mindgard's AI Security Expert subscription gives a team direct access to a specialist. Mindgard training includes instructor-led workshops where teams build agents, test attacks, configure guardrails, and harden defenses.^{[6][7][8][9][10][13]}

A US insurance brokerage used Mindgard to secure nine production AI models. Mindgard secured Fortune 500 design partners. Mindgard expanded its leadership with CEO James Brear, Head of Research Aaron Portnoy, and Offensive Security Lead Rich Smith. Mindgard lists SOC 2 Type 2. Mindgard says 86% of its staff are on the R&D team and 38% hold PhDs.^{[1][5][6][7][8][9][10][13][14]}

Mindgard on the record

Sources

Dreadnode full record

[1]Dreadnode v2 security agent infrastructure platformDreadnode · dreadnode.io · published March 24, 2026 · Tier A
[2]dreadnode.ioDreadnode · dreadnode.io · undated page, accessed November 6, 2026 · Tier A
[3]Dreadnode — AI Infrastructure for CybersecurityDreadnode · undated page, accessed November 10, 2026 · Tier A
[4]The Dreadnode PlatformDreadnode · undated page, accessed November 10, 2026 · Tier A
[5]Pricing | DreadnodeDreadnode · undated page, accessed November 10, 2026 · Tier A
[6]Credits | DreadnodeDreadnode · undated page, accessed November 10, 2026 · Tier A
[7]Secrets | DreadnodeDreadnode · undated page, accessed November 10, 2026 · Tier A
[8]Organizations | DreadnodeDreadnode · undated page, accessed November 10, 2026 · Tier A
[9]Workspaces | DreadnodeDreadnode · undated page, accessed November 10, 2026 · Tier A
[10]Projects | DreadnodeDreadnode · undated page, accessed November 10, 2026 · Tier A
[11]Beyond the Terminal: Offensive Security Evals for Embodied Reasoning | DreadnodeDreadnode · undated page, accessed November 10, 2026 · Tier A

Mindgard full record

[1]Mindgard - AI Red Teaming & AI Security SolutionsMindgard · undated page, accessed September 27, 2026 · Tier A
[2]About Mindgard: Offensive AI Security CompanyMindgard · undated page, accessed September 27, 2026 · Tier A
[3]Mindgard raises 30 million to protect ai systemsSecurityWeek · securityweek.com · published August 12, 2026 · Tier B
[4]AI Security Software: Advanced AI Testing Tools - MindgardMindgard · undated page, accessed October 26, 2026 · Tier A
[5]AI Security Services: Secure Agentic AI Systems - MindgardMindgard · undated page, accessed October 26, 2026 · Tier A
[6]AI Red Teaming Services for LLMs and Agents - MindgardMindgard · undated page, accessed October 26, 2026 · Tier A
[7]AI Penetration TestingMindgard · undated page, accessed October 26, 2026 · Tier A
[8]AI Security Advisory Services for Enterprise AI - MindgardMindgard · undated page, accessed October 26, 2026 · Tier A
[9]AI Security Consulting Services for Enterprise AI - MindgardMindgard · undated page, accessed October 26, 2026 · Tier A
[10]AI Security TrainingMindgard · undated page, accessed October 26, 2026 · Tier A
[11]Mindgard Raises $30M Series AMindgard · published August 13, 2026 · Tier A
[12]Mindgard Unveils First Autonomous AI Security Platform To Identify And Fix Exploitable VulnerabilitiesBusiness Wire · published March 17, 2026 · Tier A
[13]Offensive Security Tool for AI Systems - MindgardMindgard · undated page, accessed November 10, 2026 · Tier A
[14]Mindgard vs Prompt SecurityMindgard · undated page, accessed November 10, 2026 · Tier A
[15]AI Governance & ComplianceMindgard · undated page, accessed November 10, 2026 · Tier A

Spot a factual error? Request a correction.