Mindgard
Attacker-style reconnaissance and red teaming for AI apps and agents
Overview
Mindgard is an AI security platform spun out of Lancaster University research and headquartered in Boston and London. It runs reconnaissance and automated red-team attacks against chatbots, AI applications and agents, maps findings to OWASP and MITRE ATLAS, and adds shadow-AI discovery and runtime protection. It raised a $30 million Series A in August 2026.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 14 Oct 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
Use cases
What you can produce with Mindgard
- Reconnaissance of an AI app or agent: guardrail probing, system-prompt extraction and tool discovery
- Automated red-team tests for prompt injection, jailbreaks and encoding attacks, with ranked findings
- Findings mapped to the OWASP Top 10 for LLMs and MITRE ATLAS, with remediation guidance
- Repeat testing from CI/CD pipelines, a CLI, a Python SDK or Burp Suite
- Shadow-AI discovery and an AI bill of materials across the organisation
- Runtime protection for deployed AI systems
- Human-led AI red-team assessments from Mindgard's services team
Mindgard's strength is research you can check: it has published disclosures of real flaws in products from Google, OpenAI and xAI, and that attack knowledge feeds a platform that maps a target's guardrails and tools before attacking it. Integration through a CLI, Python SDK, CI/CD and Burp Suite makes repeat testing practical.
Pricing is unpublished and sales-led, and the tool assumes a team that can act on findings. A strong pick for organisations with AI in production; overkill for a single hobby chatbot.
What we found
Mindgard is an offensive security platform for AI systems, built on more than a decade of research by Peter Garraghan at Lancaster University. The company was set up in 2022 as Mindgard Ltd, registered in London, and now describes itself as headquartered in Boston and London. Its funding has grown in steps: a £3 million seed from IQ Capital and Lakestar in September 2023, an $8 million round led by .406 Ventures in December 2024, and a $30 million Series A led by Album VC, with Karma Ventures joining, announced in August 2026. James Brear is chief executive; Garraghan is chief science officer.
The platform works the way an attacker would. A reconnaissance stage maps what a chatbot, application or agent can do, probing its guardrails, extracting system prompts and discovering the tools it can call. Red-teaming then runs prompt-injection, jailbreak and encoding attacks, and its documentation maps findings to the OWASP Top 10 for LLMs and MITRE ATLAS with remediation advice. Tests run from the web console, a command-line tool, a Python SDK, CI/CD pipelines or Burp Suite. Newer modules add shadow-AI discovery, runtime protection and compliance reporting, and a services arm sells human-led AI red-team assessments.
Mindgard's marketing leans on its own research. It reports 156 vulnerabilities submitted to AI vendors, 132 accepted and 98 patched, with public write-ups on flaws in Google Antigravity, OpenAI Sora, the Zed editor and xAI's Grok. Those are the company's counts, but the individual disclosures are published and checkable, which is more than most competitors offer.
No prices are published and the pricing address simply returns the home page; buying starts with a demo. The documentation links to a sandbox sign-up, but its terms are not stated. Mindgard suits security teams whose AI is already in production and who need repeatable attack evidence for auditors. A hobbyist testing one chatbot will find open-source tools such as Garak or Promptfoo a cheaper start.
Compiled by RECATOOLS from public sources: vendor documentation and pricing, code repositories and published reporting.
About this listing
This entry was compiled from publicly available data including Mindgard's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Mindgard unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Mindgard directly →
Spotted something out of date? Suggest an update →
Alternatives to Mindgard
More in Security & Safety
