About Astrix

Building a More Structured Approach to Threat Modeling

Astrix is a threat modeling product from Cognexaa. It helps security and engineering teams examine a system's architecture for potential threats while it is still being designed, using established methodologies, and it leaves the decisions with the people who understand the system.

  1. Security designThe architecture your teams are planning or changing
  2. Structured threat analysisWhere Astrix works: modeling the design and examining its threats
  3. Engineering decisionsMitigations your teams own, track and build
Where Astrix sits: between the design and the engineering decisions it leads to.

Who we are

A threat modeling product, built and operated by Cognexaa

Astrix is the threat modeling platform from Cognexaa, a company that builds software for security and governance teams. Cognexaa develops and operates Astrix, and acts as the data processor for the organizations that use it.

Astrix has one job: helping teams reason about security threats in the design of applications and systems. It is not a general security suite, and this site describes only what it does today.

Product
Astrix Threat Modeling Platform
From
Cognexaa
Delivered as
A web-based service hosted in the EU on enterprise-grade cloud infrastructure, with dedicated deployments available on request
Contact
sales@cognexaa.com

The problem we aim to address

Security decisions are made in design, often without a structured review

Threat modeling is a well-established practice, but in many organizations it is hard to do consistently. These are the challenges Astrix is designed around.

  • Design moves faster than reviewArchitectures change with every release, and a design reviewed once can quickly fall out of date.
  • Systems are complexMany components, data flows, trust boundaries and external dependencies make it hard to see where threats concentrate.
  • Expertise is scarceStructured threat modeling often depends on a few experienced practitioners, so it does not scale to every team.
  • Findings get lostThreats recorded in slides or documents are hard to communicate, prioritize and follow through to a fix.

Why Astrix exists

To make structured threat modeling practical for real engineering teams

Astrix is designed to make threat modeling something teams can do on every significant design, not only when an expert is available. It starts from the artifacts teams already maintain and applies a recognized methodology consistently. Findings are expressed in standard terms that both security and engineering understand, and they are connected to the places where work gets done.

We also use the product ourselves: Cognexaa threat-models the Astrix platform with Astrix.

Our approach

Six principles behind the product

Each principle below is reflected in how Astrix works today.

  1. Start from the real design

    A threat model is only as good as its picture of the system, so Astrix builds it from current artifacts rather than redrawn diagrams.

    In practiceImport from API definitions, infrastructure-as-code, diagrams, repositories or a description.

  2. Use established methods, not a black box

    Astrix applies recognized methodologies and standard references rather than a proprietary scoring scheme.

    In practiceSTRIDE, PASTA, VAST, SAIF and MAESTRO, with CVSS, CWE and MITRE ATT&CK.

  3. Keep people in charge

    Threat modeling depends on context only the team has: the system's environment, its assumptions and its risk appetite. Astrix assists; people decide.

    In practiceAI-assisted output is advisory and reviewed by your users; a built-in rule engine can run without an AI provider.

  4. Make findings actionable

    A threat matters only if someone can act on it, so each finding is specific, prioritized and connected to delivery.

    In practiceMitigation guidance, and export to Jira, GitHub Issues or Azure DevOps.

  5. Treat design data as sensitive

    A threat model is a map of how a system could be attacked, and it deserves protection to match.

    In practiceA dedicated database per customer and no training of AI models on customer data.

  6. Be clear about scope

    We describe what Astrix does today, and where your team's judgment or other security practices are still needed.

    In practiceAutomation levels and scope notes on the features page.

What we focus on

A deliberately focused scope

Our focus

  • Threat modeling of application and system architectures
  • Security architecture analysis: components, data flows and attack paths
  • AI and agentic system designs, through AI-specific frameworks
  • Communicating and tracking findings between security and engineering

Not what Astrix does

  • Penetration testing or scanning of running systems
  • Replacing code review and application security testing
  • Security operations and monitoring
  • Certifying an architecture or your compliance

Threat modeling complements these activities. See where Astrix fits in the security lifecycle.

Who we aim to support

Designed for the people who shape system security

  • Security architects

    Who need a consistent way to reason about very different designs.

  • Application security teams

    Who need to review more systems than their headcount allows.

  • Software and application architects

    Who want security considered while the design is still flexible.

  • Engineering teams

    Who need security findings in a form they can act on.

  • Cybersecurity consultants

    Who run repeatable assessments for several clients.

  • Security leaders

    Who need visibility into design-level risk across the organization.

For scenario-by-scenario detail, see threat modeling use cases.

Security and trust

Holding sensitive design data responsibly

Customers share detailed descriptions of their systems with Astrix, so we treat its own security as part of the product.

  • Isolated: a dedicated database per customer organization.
  • EU-hosted: on enterprise-grade cloud infrastructure.
  • Encrypted: in transit and at rest.
  • Not used for training: customer data does not train AI models.

See platform security, with status labels.

What makes Astrix relevant

A focused tool for a specific job

Astrix's relevance comes from its focus: a dedicated analyzer for each of five methodologies, including two written for AI systems, a model built from real design artifacts, and findings that carry standard references and reach engineering trackers. Whether that fits your process is best judged on one of your own systems.

Looking ahead

We believe threat modeling should be a routine part of designing software, as normal as code review, rather than an occasional exercise for specialists. Astrix is our contribution to making that practical.

Explore Astrix for your threat modeling needs

Discover the current Astrix capabilities and assess how the tool may fit your application security and architecture review process.

sales@cognexaa.com