Wingman-AI

Wingman-AI is a self-hosted AI agent platform for developers and teams handling coding, research, operations or support tasks that need ongoing context. It coordinates agents with distinct roles and retains their sessions across clients.

• Specialized agents: A lead agent can delegate work to subagents, each with its own workspace, configuration and session store.
• Persistent conversations: Sessions live in the gateway and carry across devices, with access through the terminal, browser chat interface and VS Code extension.
• Recurring workflows: Scheduled routines handle repeatable work, while webhooks let external systems trigger agents.
• Message routing: Bindings direct incoming messages to a designated agent, and rooms can request parallel responses from several agents.
• Custom tools and behavior: Reusable skills, MCP tools and hooks connect agents to external systems and support automation around tool calls; tools can also present interactive components in the browser interface.

The CLI connects to a local gateway by default, with an optional local execution mode for isolated runs. A macOS desktop companion can act as a node that runs system commands and displays notifications. Wingman-AI is open source under the MIT license.