Agentdesktop 101: Governing AI Tools on Every Developer Laptop

AI coding and agentic tools are spreading across enterprise developer environments, bringing new challenges around visibility, access, credentials, and policy.

Agentdesktop is an open-source visibility, policy, identity, and management daemon for AI tools across your desktop fleet. In this introductory webinar, we’ll show how it helps teams understand what’s running on developer machines and apply consistent controls across different tools.You’ll see how agentdesktop can help you:

• Discover AI tools and resources: Inventory AI developer tools, MCP servers, skills, and models across your fleet.
• Manage configuration and policy: Define policies centrally and reconcile them into supported tools like Claude Code, Claude Desktop, Codex, and OpenCode.
• Control access and identity: Enroll devices, associate tools with users, and apply local execution policies.
• Use short-lived credentials: Provide credentials to AI tools without distributing long-lived provider API keys to developer machines.
• Gain visibility into usage: Collect session and tool-use events with user, device, and client context.

We’ll walk through agentdesktop on an individual machine, show how it works in centrally managed fleet mode, and explore how it works alongside an inference gateway such as agentgateway. Whether you’re taking stock of AI tools across your organization or ready to put consistent controls around them, this session will give you a practical introduction to agentdesktop.