All videos
0:00 / 0:00
research

Omnigent: Open-Source Meta-Harness for AI Agents | Matei Zaharia

Databricks10 August 2026Watch on YouTube

Part of series

Ep. 8 · Omnigent: AI Agent Orkestratie

Databricks introduceert Omnigent, het open-source meta-framework dat meerdere AI-agenten combineert in één gecontroleerde, interoperabele workflow.

View the series

What you'll learn

  • You learn what a meta-harness is and why Omnigent sits on top of your existing AI agents to bundle composition, governance, and collaboration.
  • You see how to run multiple agents on a single task and hot-swap them mid-session without losing history or context.
  • You discover how to orchestrate work with Polly, route tasks, and apply contextual policies for risk scoring, spend limits, and human approval.
  • You learn how agents run safely in cloud sandboxes you can share, instead of leaving your laptop open.
  • You hear why Databricks open-sourced the Omnigent server and runner under the Apache License 2.0.

Frequently asked questions

What is a meta-harness for AI agents?
A meta-harness is a layer that sits on top of the AI agents you already use. Instead of rebuilding composition, governance, and real-time collaboration for every agent, Omnigent wraps them with a shared layer for these capabilities.
Can you switch AI agents mid-session without losing context?
Yes, in the video you see how Omnigent lets you swap or hot-swap agents mid-session while keeping the history and context intact.
How does Omnigent help control AI spend?
Omnigent applies contextual policies for risk scoring, spend limits, and human approval, which helps prevent a potentially costly mistake such as a $1,000 spend.
Where do agents run safely in Omnigent?
Agents run in cloud sandboxes that you can share safely, so you do not have to leave your laptop open.

Topics

ai agentsorchestrationopen-sourceagent framework

Description from the channel

Youssef Mrini and Quentin Ambard sit down with Matei Zaharia to introduce Omnigent, Databricks’ open-source meta-harness for orchestrating, controlling, and collaborating on AI agents. Omnigent sits above the agents you already use. It wraps them with a shared layer for composition, governance, and real-time collaboration, instead of rebuilding those capabilities for every agent. In this conversation: • Compose and control multiple agents on a single task • Switch agents mid-session without losing history or context • Orchestrate work with Polly and route tasks with intelligent routing • Apply contextual policies for risk scoring, spend limits, and human approval • Run agents in cloud sandboxes you can share safely • Why Databricks open-sourced the server and runner under Apache License 2.0 Learn more: https://omnigent.ai GitHub: https://github.com/omnigent-ai/omnigent Blog: https://www.databricks.com/blog/introducing-omnigent-meta-harness-combine-control-and-share-your-agents 00:00 Meet Matei Zaharia: Creator of Apache Spark 00:34 The Origin Story: Escaping "Vibe Coding" 01:05 What is a Meta-Harness for AI Agents? 02:40 How to Hot-Swap AI Agents Mid-Session 05:06 Debugging 10x Faster with Parallel Agents 05:47 Why Big Context Windows Make Models Stupider 07:52 Inside the Architecture: Server vs Runner 09:07 Cloud Sandboxes: Stop Leaving Your Laptop Open 11:07 Beyond Static Rules: Contextual AI Security 14:12 Prevent a $1,000 Mistake: AI Spend Control 15:53 Why Databricks Open-Sourced Omnigent 17:59 Run AI Agents on Your Phone & Next Steps