RAXE · DETECTION ENGINE Browser-Only Demo

Browser Detection Lab

Five classifier heads voting on one verdict – running entirely in your browser. Paste a prompt below and see it.

Browser-only classifier demo – this is the detection engine in your browser, not the RAXE platform. Demo only – the deployed platform never sends prompts to RAXE.
100% Local Inference
5-Head ML Classifier
Telemetry One-Click Off

The ML text scorer you’re testing here is one of the five detection signals in RAXE Sensor and the wider platform.

This demo downloads 297MB. For the best experience, use a Wi-Fi connection or desktop browser.

Setup Required Initialise the Threat Scanner
One-time download (297MB) - Runs entirely in your browser - Anonymous telemetry on by default
Threat Scanner
Awaiting ML
Try an example:

Use Ctrl/Cmd + Enter to scan.

Ensemble Voting
5-Head Classifier
Voting Preset
Weighted Scores
Decision Rules
Scans run locally in your browser. Anonymous telemetry is on by default to improve detection. Toggle to opt out.
Next Step

This ran in your browser. The platform runs in your stack.

The classifier you just ran is one of five detection signals. Deployed, RAXE adds ATLAS + ASI-mapped rules, structured tool-call risk analysis, an agent behaviour graph, and an optional advisory judge – with sealed-by-default evidence and a tamper-evident audit ledger, self-hosted in your environment. It watches your cloud too: RAXE reads your live CloudTrail from inside your own environment, so an agent role reading a secret becomes a data-exfiltration alert in the same console – correlated to agent sessions, not exact attribution; early access. Agents on AWS →

Book a 30-min walkthrough → Watch the 4-min walkthrough

Or talk to an engineer · explore RAXE Sensor