Hands-on labs for security teams and consultancies. Your people exploit and fix vulnerabilities in live LLM apps, RAG systems and AI agents. You assign the work and see who has passed.
Team is $449 per seat per year, with a 4-seat minimum, excluding VAT/GST and other charges.
Team skill matrix. Sample data, not a real team.
Product teams ship chatbots, retrieval systems and agents with tool access, and clients now put them in pentest scopes. Most security staff have never tested one, and there is no shared method for reviewing them.
Train your existing testers on live AI targets, certify the ones you put in front of clients, and show which of your bench can take the work.
For AppSec and product security teams whose company ships chatbots, retrieval systems or agents. Engineers learn to test them; developers learn to prevent the same flaws.
Labs are live, deliberately vulnerable LLM apps, RAG systems and agents.
A lab completes when the platform checks the evidence: the exploit landed, or the fix held against the attack.
You get: A graded completion record for every lab each person finishes.
Each OWASP topic has a written lesson and a concept video ahead of its labs.
You get: One set of paths for everyone, across all 200+ labs.
Sample data, not a real team.
Decide who trains on what and by when, then see who is done.
Assign a learning path or single labs to people, groups or roles, with a due date.
You get: An assignment list with done-of-total for each path, against its due date.
Sample data, not a real team.
Coverage is built from graded labs, so it reflects work people have done.
Each person against the categories of the OWASP LLM, Agentic and MCP Top 10s, with the team's biggest gaps listed.
You get: A coverage view per person and a gap list for the team.
Hand coverage to a security review, an auditor or a client.
You get: A coverage report as PDF or CSV, and an evidence PDF per framework.
Sample data, not a real team.
GSCP and GSCE are hands-on exams against a live AI application, reviewed by an examiner.
Assign exams to people on the team and follow each sitting through to the result.
You get: One screen showing who holds an exam, where each sitting stands, and which credentials were issued.
Every credential an examiner releases has its own public page.
You get: A certificate per person with a verification URL a client or auditor can open.
There was a clear need for AI-native sandboxed labs, and GenAI Security Lab meets that need well. The learning paths are structured and easy to follow, and the labs reflect realistic application scenarios we encounter in day-to-day work. Overall, this platform is an effective way to learn current LLM, agentic, and MCP vulnerabilities.
GenAI Security Lab does a good job combining structured OWASP-based training with practical, scenario-driven labs and automated verification. What stands out is how it connects offensive techniques like MCP tool schema poisoning and cross-tool data exfiltration back to the actual vulnerable backend code, the root causes, and practical defenses — including code-level fixes and blast-radius containment. For pentesters, it helps you build more realistic test cases, trace the full attack chain, confirm the real security impact, and produce evidence-backed findings with clear, actionable remediation.
For a squad training together, with the admin tooling to run it.
For organisations rolling training out across security and engineering, with procurement to satisfy.
Buying for yourself? Individual plans are on the .
| Features | Team $449 per seat per year | Enterprise From $999 per seat per year · 5+ users |
|---|---|---|
| Training | ||
Labs | Every lab | Every lab |
Full curriculum | ||
Progress tracking | ||
| Certification | ||
GSCE certification attempt | 25% off | GSCP + GSCE for every seat |
| Team management | ||
Centralized billing & seat management | ||
Assign labs & learning paths | ||
Groups & role-based access | ||
Coverage insights | ||
Usage & engagement analytics | ||
Compliance & certification reports | ||
Audit log | ||
SSO / login policy | Email/password, Google & GitHub, passwords-off switch, custom session timeout & allowed domains | Team's sign-in policy, plus SAML SSO and SCIM on request |
| Add-ons | ||
Instructor-led training access | Bought separately | Bought separately |
Custom labs access | Bought separately | Bought separately |
| Support | ||
In-app feedback & support | ||
Support channel | Email + Slack | Dedicated support |
These sit alongside any plan and are quoted separately — they are not part of a subscription tier.
A private lab built around one of your own applications — your stack, your data shapes, one attack technique.
Instructor-led days taught by the founder, Aditya Agrawal, run privately for your team on dates agreed with you.
Seats, tracking, certification and how to start.
Team is $449 per seat per year, with a 4-seat minimum, so the smallest invoice is $1,796 a year. Every seat is billed at the same rate, with no base fee. Seats are added from the billing page and prorated when added mid-term. A seat is for one named person who trains, and can be reassigned when someone leaves. Team Admins, Billing Admins and Read Only Users take no seat and are not billed. Team costs the same in every country: it is priced in US dollars, excluding VAT/GST and other charges, which are added at checkout where they apply. Enterprise starts at $999 per seat per year, from 5 users.
An admin assigns learning paths or single labs to people, groups or roles, with a due date. Labs are graded by the platform, so nobody reviews attempts by hand. The workspace shows who has finished each assignment and each person's coverage across the OWASP LLM, Agentic and MCP Top 10s, and exports a coverage report as PDF or CSV.
GSCP and GSCE are hands-on exams against a live AI application, and an examiner reviews every sitting before a credential is issued. Each certificate has a public verification URL. On Team they are bought separately (GSCP $199, GSCE $399), at 25% off on every Team plan, since all of them are yearly. Enterprise includes both for every seat.
Centralized billing and seat management, the ability to assign labs and learning paths, groups with role-based access, coverage insights, usage and engagement analytics, compliance and certification reports, an audit log, and a login policy covering sign-in methods, allowed domains, and session timeout (SAML SSO and SCIM are Enterprise features, set up on request). It's the console that lets you see exactly who's covered and who isn't.
Book a demo or email [email protected], and we set up the team workspace with you. You can also run a pilot on the full Team plan for up to 30 days, with no card. Anyone on the team can try the free labs with an individual account first.
Our security posture, subprocessors, data retention and responsible-disclosure policy are on the Trust & security page.
In 30 minutes: assign a lab, inspect one graded fix, and export the resulting coverage report.