Omnigent — a meta-harness for building and running AI agents

Omnigent is an open-source meta-harness that acts as a common layer for composing, governing, and collaborating on AI agents — coding and otherwise — on top of the harnesses you already use.

Key capabilities

• Local runners: Runners started on your laptop have direct access to your machine's tools, files, and credentials.
• Agent registration: Each registered agent is stored in a top-level acp: block under ~/, and agents can be imported from an existing OpenClaw/acpx registry.
• Multi-agent collaboration: Agents stream events back to the server over WebSocket and can tell collaborators apart mid-conversation.
• Sub-agent delegation: /investigate delegates read-only work (debugging, audits, code understanding) to sub-agents and synthesizes findings.
• Model and provider support: Supports Ollama (local), Codex / OpenAI agents, OpenRouter, and more; for Claude Code, point at OpenRouter's Anthropic-compatible endpoint.
• Approvals and policies: Inbox approvals resume the gated tool call, and default policies created via the API take effect on sessions.
• Appearance customization: Pick a color theme (Omnigent, Dracula, GitHub, Catppuccin, Gruvbox), a terminal theme, and independent UI and code editor/terminal font size and family.

Platforms

Available on Android, iOS, and macOS.