World's First Autonomous AI Agent Operating System benni-os.net

The Operating System
where AI agents
live & evolve.

Benni OS is an operating system designed from the ground up for AI agents to live, think, plan, execute, and evolve — independently. Persistent memory. Multi-agent coordination. Native tooling. Governance layers. Self-improvement capabilities.

Explore the System GitHub MONOMO IDE
JARVAS-2 online_active
Benni Control Plane running
6 Swarm Formations deployed
4 live content channels
Open Source first
OS Manifesto

Designed for agents that don't wait for instructions.

Every OS before Benni OS was built for humans — files, folders, GUIs, all metaphors for human cognition. Benni OS is different. It is the first OS where AI agents are first-class citizens.

LIVE
Persistent agents with memory
THINK
Multi-model reasoning layers
PLAN
Autonomous strategic objectives
EXECUTE
Swarm formations dispatched
EVOLVE
Self-improving skill system
Core Infrastructure

Products Built & Running

Every component of Benni OS is production-grade, running in real time. Not a prototype — a live system operating at the apex of AI technology.

● LIVE
SA-93 · Core Agent Runtime
JARVAS-2
The autonomous agent runtime powering the entire Benni OS. Multi-layer architecture — Orchestrator, Swarm Dispatcher, Content Pipeline, Revenue Loop, Memory Agent, Control Plane and Token Economy. Status: online_active.
L-1Orchestrator Core● online
L-2Swarm Dispatcher (SA-96)● online
L-3Content Pipeline (SK-50)● online
L-4Revenue Loop (SK-95)● online
L-5Memory Agent (SK-60)◐ standby
L-6Control Plane (SK-90)◐ standby
L-7Token Economy (SA-82G)● online
● LIVE
SK-90 · Control Plane
Benni Control Plane
Mission control for the entire Benni OS ecosystem. Real-time monitoring of agents, formations, objectives, memory and revenue loops. Connected to GitHub MCP and JARVAS-2 MCP.
Real-time agent status dashboard
GitHub MCP integration live
Objective & run management
Snapshot & state recovery
PythonMCPGitHubJARVAS-2
⚙ IN BUILD
SK-20 · Native IDE
MONOMO IDE
The IDE built to be the body of Benni OS. Born agent-native, breaking every barrier of market IDEs by default. Live formation monitor, built-in MCP tooling and deep JARVAS-2 integration.
Agent-first file system
Live Swarm Formation monitor
Native MCP tooling — no extensions
Persistent agent memory surfaces
ElectronTypeScriptLSPWASM
SK-20 · Native IDE

MONOMO IDE

The ideal body for Benni OS. An IDE born to run AI agent projects — breaking all market barriers by default. Where VS Code was built for humans writing code, MONOMO is built for agents writing their own future.

  • Agent-First Workspace
    File system, terminal, and sidebar designed around agent runs. Agents are processes, not plugins.
  • Live Formation Monitor
    All 6 Swarm Formations (ALPHA–FOXTROT) live in the sidebar. Dispatch, cancel, inspect from inside the IDE.
  • Native MCP Tooling
    GitHub MCP and JARVAS-2 MCP wired natively — no extensions needed. The IDE understands agent protocols.
  • Persistent Agent Memory
    Memory namespaces, cognitive stores and episodic logs surfaced in the editor. Context never disappears.
agent_runtime.py
formation.yaml
benni-os
agent_runtime.py
jarvas2_core.py
swarm_dispatch.py
formations
alpha.yaml
bravo.yaml
memory
cognitive.db
1# BENNI OS · Agent Runtime v2.0
2
3from jarvas2 import Orchestrator, MemoryAgent
4from benni_os.swarm import Formation, dispatch
5
6class AgentRuntime:
7 def __init__(self):
8 self.orc = Orchestrator("benni-os")
9 self.mem = MemoryAgent(persistent=True)
10 self.formations = []
11
12 async def boot(self):
13 await self.orc.initialize()
14 # dispatch ALPHA & BRAVO by default
15 await dispatch(Formation.ALPHA)
16 await dispatch(Formation.BRAVO)
17 return "online_active"
● JARVAS-2 Python 3.11 ALPHA ● BRAVO ● MONOMO v0.1-alpha
SA-96 · Army Protocol

Swarm Formations

Autonomous agent dispatch. Handoff via artifact_id. Max 2 active formations simultaneously. Each formation is a specialized agent team with a single mission.

ALPHA
Launch
Coordinates product and channel launches. Activation sequence, copy, and distribution pipeline.
● active
BRAVO
Build & Ship
Development and delivery of software, automations and systems. Code → test → deploy.
● active
CHARLIE
Growth
Channel growth strategy, audience building and organic content distribution.
◐ standby
DELTA
Research
Deep market, technology and opportunity research. Direct feed to Orchestrator.
◐ standby
ECHO
Health
System monitoring, project health, alerts and performance diagnostics.
◐ standby
FOXTROT
Revenue Zero
Emergency mode for immediate revenue generation with minimal resources.
◐ standby
Open Source First

Open Source Projects

Built in the open. Every core component of Benni OS is open source — because the future of AI infrastructure belongs to everyone.

benni-os

The official Benni OS website and system documentation. Landing page, manifesto, products, formations and datacenter vision.

HTML/CSS/JS
Public
benni-master-os-skills

The complete skill system (SK-00 to SK-96). Orchestrator, Revenue Loop, Army Protocol, and all formation definitions powering JARVAS-2.

YAML + Python
Public
benni-control-plane

Mission control dashboard. Snapshot system, approval workflows and real-time agent health monitoring connected to JARVAS-2 MCP.

Node.js
Public
monomo-ide

The native IDE for Benni OS. Agent-first workspace with built-in MCP tooling, formation monitor and persistent memory surfaces. In active build.

Electron + TS
Building
jarvas2-mcp-server

MCP server exposing JARVAS-2 tools — run_task, queue_task, save_memory, get_snapshot — to any MCP-compatible AI client.

Python
Public
content-pipeline-ai

Automated content generation pipeline for YouTube. Script → TTS → video → thumbnail → metadata → publish. Fully orchestrated by JARVAS-2.

Python
Public
Next Strategic Objective

Benni AI Datacenter

The ultimate objective: building our own AI datacenter to carry the most powerful operating system in the world — Benni OS. Full self-hosted inference. Persistent agent memory. Private multi-GPU cluster. Zero dependency on external cloud providers.

Self-hosted inference — Local LLMs, zero latency, full sovereignty
Persistent vector memory — Agent memory that never expires
Multi-GPU cluster — Power to run the full swarm simultaneously
Private & sovereign — No Big Tech dependency
Global reach — Castanhal, Pará, Brazil → World
Benni AI Datacenter — Phase 1 RackDC-CAST-01
JARVAS-2 Runtime Node● running
Benni Control Plane Server● running
Content Pipeline Server● running
Local LLM Inference [Hermes]◐ cloud-fallback
GPU Node 1 [RTX 4090]◎ planned
GPU Cluster — Multi-node◎ planned
Persistent Vector DB◎ planned
Infrastructure Progress35%
Revenue → Hardware FundActive
Current System Uptime99.2%
SK-50 · Content Engine

YouTube Channels

Automated content ecosystem powered by Benni OS pipelines. AI-generated scripts, TTS, video, thumbnails and publishing — fully orchestrated by JARVAS-2.

Cerebrox
AI, autonomous agents and open-source technology. The Benni OS content hub.
PT-BR · AI & Tech
Adam Invests
Financial education and investment psychology. Long-term thinking, AI-generated.
EN · Finance
Doomsday POV
Extreme scenarios, survival psychology and immersive AI storytelling.
EN · Survival
No Panic Finance
Finance without the drama. Behavioral economics and evidence-based investing.
EN · Psychology & Finance
// join the system

The future of AI is
autonomous by design.

Benni OS is not a product. It is an operating system for the age of intelligent agents. Follow the build, contribute to the open source, or connect directly.

Follow on GitHub Watch Cerebrox