Pragatix AI Firewall
AGAT Software · Cybersecurity & IT
Certification per AWS Marketplace.
Evidence tier Source Confirmed · 4 captures on record
What the publisher says
As described on AWS Marketplace.
Pragatix AI Firewall is a security and governance layer designed for organizations that want to adopt public AI tools without compromising control, security, or compliance. It enables companies to mitigate the risk of utilizing public AI services by implementing AI policies that are tailored to user intent and the sensitivity of the data involved.
The AI Firewall provides three layers of protection: Usage, Agent, and Model.
Show the rest of the publisher’s description (4 more lines)
The Usage layer functions as a real-time proxy governing access to public AI services such as ChatGPT, Copilot, and Gemini.
The Agent layer controls AI agents and tools executed via MCP servers, enforcing policy on tool access and behavior.
The Model layer includes a built-in Risk Engine that evaluates AI models using source intelligence and static and dynamic vulnerability assessments.
Across all layers, the platform mitigates AI risks by applying guardrail policies based on users and groups, classifying data sensitivity, and providing centralized visibility and analytics. Coverage across browser extensions, network proxies, and API integrations ensures protection against key OWASP AI threats, including prompt injection, data leakage, and model abuse, across both enterprise and shadow AI usage.
Highlights
Highlighted by the publisher on AWS Marketplace.
Three-Layer AI Defence Secures AI usage, agents, and models through a unified governance and security framework.
Real-Time AI Governance Applies risk-based policies to control access to public AI services and AI agents while classifying data sensitivity.
OWASP AI Risk Protection Mitigates common AI threats such as prompt injection, data leakage, and model abuse with full usage visibility.
Agent build and provenance
See the full provenance
The layer-by-layer build, the evidence behind each claim, the risk basis and the cross-marketplace links are open to any account. Some rows are disclosed, some the source leaves Unknown; a free account shows you which.
Plans and pricing as listed
2 listed- Users
- Users
Refund terms
As stated by the publisher on AWS Marketplace.
Refunds are applicable in accordance with these two scenarios: 1.The product does not accomplish what has been documented, whether on the website or in any official AGAT Documentation. All information about the solutions is available on www.businessgpt.pro . 2.AGAT`s technical support team has updated the customer that the issue will not be fixed in a reasonable time frame.
Sources
Publisher resources
1 linkLinked repositories
Unknown means this listing does not publish a repository. It is not a statement that the code is closed, and a linked repository is not a claim that the publisher wrote it: the registry computes that relationship privately and does not publish it.
Evidence risk is the share of the build you cannot see before you deploy, not a security rating. Sign in to see the layer-by-layer basis for this band.