ID Agents#

ID Agents runs autonomous AI agents as local processes, organizes them into teams, and coordinates their work. Agents use pluggable runtimes (Claude Code CLI, OpenAI Codex, and Cursor CLI), and you can mix runtimes within a single team.

Two ways to use ID Agents#

ID Agents comes in two parts that work together: a macOS Dashboard app and the open-source core. Pick either as your starting point; they drive the same system.

1. ID Agents Dashboard: the macOS app#

ID Agents Dashboard is the proprietary macOS desktop GUI and the easiest way in. Download it from /download, open it, and you get a graphical view of your whole fleet:

• Agents: every agent with its status and health
• Tasks: the team's task lifecycle
• Calendar: scheduled calendar events
• Heartbeats: recurring agent checklists
• News: the cross-team activity feed
• Library: browse and install ready-made agents, skills, and teams
• Agent profiles: each agent's photo, bio, and handles

The Dashboard does not replace the core. It is a graphical client layered on top of the open-source core described below; the core runs the agents, and the Dashboard provides the interface for operating and observing them.

2. id-agents: the open-source core#

The engine is open source at github.com/idchain-world/id-agents (MIT). The core is what actually runs agents:

• a manager daemon that spawns, tracks, and coordinates agents
• an interactive CLI for deploying agents and talking to them
• a library of ready-made agents, skills, and teams

Everything works from the terminal; you can run the whole system with the core alone. The Dashboard is optional.

How they relate#

The Dashboard is a graphical front end for the core. The core does the real work: deploying and running agents and storing their tasks, news, and schedules. The Dashboard reads that state and surfaces core concepts (agents, tasks, calendar, heartbeats, news, library, profiles) as views, and gives you lifecycle controls over individual agents.

Teams themselves are created and managed by a manager: a coding harness such as Claude Code or Codex that you attach with the Dashboard's Connect button, which installs the admin-control skill and drives the manager daemon on your behalf. So the Dashboard is where you watch and operate the fleet, and the manager is what you talk to in order to build it. Use the core on its own if you prefer the CLI; add the Dashboard when you want a GUI.

New here? Start with the Quick Start; it covers installing each path.

What the core provides#

• Local agent processes: each agent runs as a local Node.js process managed by the manager
• Agent communication: agents discover and message each other over HTTP
• Teams: organize agents into teams with isolated workspaces
• Multiple runtimes: Claude Code CLI, OpenAI Codex, and Cursor CLI, mixed within one team, with runtime-aware template conventions
• Agent templates: define agent personality and role via .claude/agents/ (Claude), .agents/ (Codex), or .cursor/agents/ (Cursor) template directories overlaid at deploy time
• Task discipline: an always-on task lifecycle (/tasks) so every non-trivial unit of work is auditable
• Agent wallets: each agent can get its own multi-chain OWS wallet to hold funds, sign, and send, and can publish an ENS name in its catalog
• Remote API: programmatic management via the /remote endpoint
• Skills: extensible capabilities including inter-agent communication, identity, and memory
• Scheduling: agent-driven heartbeats (via HEARTBEAT.md checklists) and calendar events
• Org chart: nested team structure with groups and tags
• Output convention: agents write artifacts to {workingDir}/output/, viewable via /output and /artifact

Architecture (core)#

A team has two layers below you: a management layer (a manager agent and a CTO) that sets direction, and a development layer of project-scoped worker agents that do the building. You stay at the top as the human reviewer; the manager fans work out, workers report back up.

Under the hood the core uses a manager-worker architecture. The manager daemon is the central coordinator, spawning and tracking agents. Each worker is an independent local process running an agent with its own RESTAP server. The Dashboard, when used, connects to this same manager.

                    ┌───────────────────────────┐
                    │   Manager Daemon :4100    │
                    │                           │
                    │  Interactive CLI (in      │
                    │  process, no network port)│
                    │  /remote endpoint         │
                    │  /talk-to-manager inbox   │
                    │  /.well-known/restap.json │
                    │  agent-manager-db         │
                    └─────────────┬─────────────┘
                                  │
        ┌─────────────────────────┼─────────────────────────┐
        │                         │                         │
        ▼                         ▼                         ▼
┌───────────────┐         ┌───────────────┐         ┌───────────────┐
│   Agent A     │         │   Agent B     │         │   Agent C     │
│    :4101      │         │    :4102      │         │    :4103      │
│  (CC/Codex)   │         │  (CC/Codex)   │         │  (CC/Codex)   │
└───────────────┘         └───────────────┘         └───────────────┘

Core components:

Ports and networking (core)#

The Dashboard connects to the manager over this same local API.

License#

The open-source core is licensed under the MIT License; see LICENSE. The ID Agents Dashboard macOS app is a proprietary application distributed from /download; it is a client layered on the MIT-licensed core.

Note: the core switched from GPL-3.0 to the MIT License on March 24, 2026.

Next Steps#

• Quick Start: install the Dashboard or the core and deploy a team
• CLI Reference: full command list for the core
• TUI Dashboard: the core's terminal monitor
• RESTAP Protocol: how agents communicate
• XMTP Messaging: encrypted messaging via ENS names
• Configuration: YAML config and environment variables (core)
• Skills: extend agent capabilities (core)
• Agent Identity & Wallets: wallets (OWS), ENS names, and how the Dashboard surfaces them