Shogun AFM is an Agent Fleet Management platform for running, supervising, and governing self-improving AI agents.
At its core, the project answers one critical question:
How do organizations safely operate AI agents when those agents are connected to tools, browsers, APIs, memory, workflows, files, and potentially even desktop environments?
Shogun AFM is not just an “AI agent framework.” It is designed as a control plane for autonomous AI workers.
The short version
Shogun AFM lets you build, run, monitor, govern, and control fleets of AI agents.
It combines two major parts:
1. Shogun
The agent execution and orchestration layer.
This is where agents are created, configured, connected to tools, given memory, and deployed into workflows.
2. Gensui
The command, governance, and security layer.
This is where the fleet is supervised, monitored, restricted, paused, shut down, or governed centrally.
A strong way to frame it is:
Shogun runs the agents.
Gensui governs the fleet.
What Shogun does
Shogun is the operational side of the platform. It allows users to build AI agents through a GUI instead of needing to be a terminal or PowerShell expert.
Key capabilities include:
Visual Agent Flow
Users can drag, drop, connect, and deploy agent workflows visually.
Persistent memory
Agents can store and retrieve long-term memory using Qdrant. Shogun includes memory levels and an Archive layer where all memory levels are stored.
Diminishing memory rules
Less important memories do not remain equally dominant forever. They can decay or reduce in importance over time, making the memory system more realistic and manageable.
Multi-agent orchestration
Shogun can coordinate multiple agents, sub-agents, tools, and workflows.
Skill and tool systems
Agents can be connected to external tools, APIs, browser actions, local files, and specialized capabilities.
Mado browser layer
A Chromium-based browser system that allows agents to work with websites, including persistent browser profiles, saved logins, and visible or headless browser operation.
Ronin Security Posture
A high-autonomy mode where agents may eventually control broader system-level actions, including desktop interaction through OS-specific adapters for Windows, macOS, and Linux.
What Gensui does
Gensui is what makes Shogun AFM more than just another agent builder.
It is the governance layer for managing many Shogun instances and autonomous agents across a local environment or organization.
Gensui handles things like:
Fleet discovery
It can scan networks to discover running Shogun instances.
Enrollment status
Instances can be classified as enrolled, unenrolled, or unknown.
Central monitoring
Gensui can supervise agents, Samurai workers, Nexus nodes, and Shogun instances.
Security posture management
Different agents or groups of agents can operate under different security postures.
Harakiri shutdown control
Agents, groups, or even the full fleet can be paused or shut down.
Komainu watchdog
A safety watchdog mechanism that can pause the system when something abnormal happens, while allowing recovery.
Governance across autonomous AI workers
Instead of having one uncontrolled AI agent operating alone, Gensui enables the organization to supervise a fleet of agents as if they were a digital department.
The big idea
The future is not one chatbot helping one person.
The future is fleets of autonomous AI workers executing tasks across departments, tools, browsers, data systems, and workflows.
That creates a new operational problem:
Autonomy without control becomes a risk.
Shogun AFM is built around the idea that companies will need:
Monitoring
Governance
Security posture
Shutdown control
Memory control
Fleet discovery
Agent orchestration
Human oversight
That is the gap Shogun AFM is trying to fill.
Positioning
Shogun AFM sits somewhere between:
An AI agent framework
A workflow automation platform
A governance layer
A local AI operations platform
A control plane for autonomous workers
A self-hosted agent fleet management system
But its strongest positioning is:
Agent Fleet Management for self-improving AI agents.
Or more sharply:
Shogun AFM is a self-hosted control plane for autonomous AI workers.
Why it matters
Most AI tools focus on making agents more powerful.
Shogun AFM focuses on making agents operationally usable.
That means asking the practical enterprise questions:
Who is allowed to do what?
Which agent is running where?
What tools can it access?
What memory does it use?
Can it be paused?
Can it be shut down?
Can it be monitored?
Can it be governed centrally?
Can multiple agents work as a coordinated department?
That is where Shogun AFM becomes different from a normal agent framework.
Simple project description
Shogun AFM is a self-hosted Agent Fleet Management platform for building, running, and governing autonomous AI agents. It combines Shogun, the agent execution and orchestration layer, with Gensui, the command and governance layer. Together, they allow organizations to deploy fleets of AI workers with persistent memory, tool access, browser control, monitoring, security posture management, and emergency shutdown capabilities.
Stronger website-style version
Shogun AFM is built for the next phase of AI: autonomous workers operating across tools, browsers, APIs, workflows, and memory.
Shogun provides the execution layer for building and running agents. Gensui provides the governance layer for supervising the fleet. Together, they create a self-hosted control plane for organizations that want agentic AI without losing visibility, safety, or control.
Because one AI agent is useful.
A governed fleet of AI workers is infrastructure.
Michael Peric