v1.15.0
new
Network-Level AI Control, LLM Classification & Audit Dashboard
- Cloudflare Gateway integration — block unapproved AI tools at the DNS level, covering all devices (browser, native apps, mobile, CLI). Connect in Settings → Integrations.
- AI Tool Policy — new tab on Guardrails page lets admins approve/block 18 AI tools. Syncs to Cloudflare Gateway when connected, enforces via extension when not.
- LLM topic classification — define sensitive topics in plain language (e.g. 'customer financial data'). AI classifies prompts against your topics using OpenAI or Anthropic.
- User education on block — when the extension blocks a message, it now explains WHY it matters and WHAT to do instead, with per-category guidance for PII, credentials, API keys, health data, and more.
- Audit & Compliance page — new /audit dashboard with Sankey flow diagrams (team→tool), violation heatmaps, risk score histograms, donut charts, compliance framework coverage, and CSV/PDF export.
- Dashboard restructured — sidebar simplified to 7 items (was 9). Template Packs and Guidelines merged into Prompt Library (tabs + Browse Packs sheet). Analytics replaced by Audit.
- Integrations reorganized — Security & Governance, Directory & Team Management, and Developer Tools sections.
- Onboarding — setup checklist now includes optional Cloudflare Gateway step.