By Craig Jellick, VP Engineering & Co-founder of Obot AI
v0.24.0 has shipped, and we're proud to highlight Obot Sentry — a lightweight companion agent that enrolls a workstation with Obot, runs scheduled device inventory scans, and installs managed audit hooks for Claude Code, Codex, VS Code, and Cursor.
Why Obot Sentry? Teams are trying to understand what AI is doing across their organizations, but activity is scattered. Some traffic goes through an LLM gateway. Some happens in local AI clients that never cross a centralized endpoint. Some is in MCP servers, some is direct model calls. No single capture point covers all of it.
v0.24.0 adds three pieces to address that:
→ Obot Sentry — device-level audit hooks for local AI clients; first MDM path is Microsoft Intune for Windows
→ Full LLM Gateway audit logs — every model request logged with user, provider, model, token usage, duration, and encrypted content; exportable to S3, GCS, or Azure
→ AWS Bedrock, Azure, and Generic Responses providers — broader model provider coverage without a different governance model for each
Plus: composite MCP servers through GitOps, a significantly smaller MCP proxy footprint, built-in local auth, and in-app getting started guides.
📓 Full release notes · 🐙 GitHub v0.24.0
Give Obot a try - no strings, no card needed, full enterprise edition, free for 2 weeks.
Originally published on Obot AI
Top comments (0)