Omnigent Updates: Swap AI Coding Agents, Slack Integration, Policies, Automation & Usage Tracking
Omnigent keeps shipping fast, and this update covers the features that actually matter for daily use. 🚀 I walk through Omnigent as a meta-harness layer over Claude Code, Codex, Pi, and your own custom agents, letting you swap or combine harnesses without rewriting anything, while policies, guardrails, and sandboxing stay in place underneath. Along the way: the new Slack integration, a full UI tour, a live Polly demo dispatching multiple agents against a real CSV, Projects for organizing sessions, policy guardrails including PII detection and dangerous shell-command blocking, Automation for recurring runs, and how usage telemetry and per-session cost tracking actually work. ───────────────────────────────────── 🔥 WHAT YOU WILL LEARN ───────────────────────────────────── ✅ What Omnigent is: a meta-harness layer that sits above Claude Code, Codex, Pi, and custom agents ✅ Why a common agent layer matters, swap or combine harnesses without rewriting your setup ✅ How to install Omnigent and set up the new Slack integration to chat with your agentsomn ✅ Managed (Databricks Beta) vs self-hosted open source, which deployment fits your setup ✅ A tour of the updated UI: sessions, files, and the Polly orchestrator dispatching multiple agents ✅ Polly analyzing a real CSV live, with cross-vendor review before you merge ✅ Projects: grouping related sessions together in the sidebar with per-project defaults ✅ Policy guardrails in action: PII detection and blocking dangerous shell commands ✅ Quick model and harness switching mid-session ✅ Automation: scheduling recurring agent runs so you stop kicking things off by hand ✅ Usage telemetry, what it actually collects and how to turn it off ✅ Tracking per-session cost straight from the terminal with omni usage FREE the whole way through: everything shown here runs on the free, self-hosted, open source tier. Previous video: https://youtu.be/h_9ix0IJ96g?si=TQjcV6THTjVX6smf ───────────────────────────────────── ⏱ CHAPTERS ───────────────────────────────────── 00:00 Welcome and Overview 00:40 What Is Omnigent 01:11 Why a Common Agent Layer 02:21 Install and New Slack Integration 03:41 Managed vs Open Source 05:05 Polly Demo on CSV 06:37 Projects and Session Grouping 07:53 Policies and Guardrails 09:30 PII Detection Demo 12:05 Automation and Scheduling 12:41 Telemetry and Privacy Controls 13:45 CLI Sync and Cost Tracking 14:26 Wrap Up and Next Steps ───────────────────────────────────── 🔗 RESOURCES ───────────────────────────────────── ↗ Omnigent official site → https://omnigent.ai ↗ Install guide → https://omnigent.ai/quickstart/install ↗ Docs: Slack integration → https://omnigent.ai/docs/interact/slack ↗ Docs: Web UI (Projects, sessions) → https://omnigent.ai/docs/interact/web-ui ↗ Docs: Builtin policies (PII, shell guardrails, cost) → https://omnigent.ai/docs/policies/builtin ↗ Docs: Scheduled Tasks (Automation) → https://omnigent.ai/docs/build/scheduled-tasks ↗ Docs: Usage Telemetry → https://omnigent.ai/docs/deploy/telemetry ↗ Docs: Harnesses (swap/combine agents) → https://omnigent.ai/docs/build/harnesses ↗ Databricks-managed Omnigent (Beta) → https://docs.databricks.com/aws/en/omnigent/ ↗ GitHub repo → https://github.com/omnigent-ai/omnigent ↗ Full changelog → https://github.com/omnigent-ai/omnigent/blob/main/CHANGELOG.md ↗ Discord community → https://discord.gg/omnigent ───────────────────────────────────── ☕ Buy me a Coffee: https://ko-fi.com/datasciencebasics ✌️ Patreon: https://www.patreon.com/datasciencebasics ------------------------------------------------------------------------------------------ 🤝 Connect with me: 📺 Youtube: https://www.youtube.com/@datasciencebasics?sub_confirmation=1 👔 LinkedIn: https://www.linkedin.com/in/sudarshan-koirala/ 🐦 Twitter: https://twitter.com/mesudarshan 🔉 Medium: https://medium.com/@sudarshan-koirala 💼 Consulting: https://topmate.io/sudarshan_koirala #omnigent #databricks #omnigenttutorial #datasciencebasics