stack.tools

Vega Security

Agentic cyber defense platform that detects, triages, and investigates threats on security data in place.

Vega runs an agentic cyber defense loop — detection, triage, investigation, optimization — directly on customer security data without ingesting it into a SIEM, with Claude as the reasoning engine. The team deliberately works inside a single model family, matching Haiku, Sonnet, and Opus to each pipeline stage so the deepest reasoning lands on confirmed alerts while high-volume log analysis runs on lighter tiers. Customer inference is served through Amazon Bedrock, while the direct Claude API covers internal tooling and Claude Code.

vega.io
Facts
2
Receipts
3
Layers
2

Sources last verified Sep 28, 2026

Models & Providers

1 fact

Claude by Anthropic

Reasoning engine behind the agentic detection, triage, and investigation loop.

High confidence

Receipts · 2

Show quotes (2)
  • “The team chose Claude to power those agents in production.” — claude.com, Aug 2026
  • “with Claude as a reasoning engine behind it” — claude.com, Aug 2026

Source last verified Sep 28, 2026

AI Coding Tools

1 fact

Claude Code by Anthropic

Used internally for engineering work, served by the direct Claude API.

Medium confidence

Receipts · 1

Show quotes (1)
  • “the direct Claude API runs alongside it for internal tooling and Claude Code.” — claude.com, Aug 2026

Source last verified Sep 28, 2026