Skip to content
ZekaForge
Product

How the AI security assistant works

A provider-independent AI layer (Claude via Anthropic or Amazon Bedrock) that analyzes real findings.

The AI assistant turns raw scan output and logs into clear, prioritized guidance.

Provider-independent by design

ZekaForge uses an adapter layer so the same features work across providers:

  • Anthropic Claude API
  • Amazon Bedrock (Claude models), keeping data in your AWS account
  • additional providers via the same interface

A demo driver returns deterministic, clearly-labeled sample analysis when no key is configured.

What it does

  • Explains vulnerabilities in natural language
  • Prioritizes risk and suggests an order of remediation
  • Drafts executive summaries for non-technical stakeholders
  • Answers questions about a specific scan or log

Controls

Per-tenant token quotas, cost tracking, request logging, retries, and secure key storage are built in. Secret keys live in environment variables or AWS Secrets Manager — never in frontend code.

Languages

The assistant can respond in all five platform languages, so an analyst in Riyadh and a manager in Shanghai read the same finding in their own language.

分享

相关文章

开始使用 ZekaForge

创建免费账户,几分钟内体验 Device Cloud 与 Cyber AI。