Project Laputa
The Self-Earning Intelligence Layer
An autonomous floating intelligence that observes the world, discovers opportunities, deploys agents, captures revenue, and self-evolves — the self-sustaining Decision OS.
Observe → Judge → Execute → Earn → Evolve
An OS That Earns Its Own Revenue
Laputa is not a tool you operate. It is a floating intelligence structure — each layer observes, judges, acts, earns, and evolves.
When this loop closes, the OS sustains itself — self-driving revenue.
Five Layers of Floating Intelligence
Observe markets, regulations, news, customer data
Bundle and score proof before any action
ROI, risk, priority, responsibility boundaries
Deploy agents to deliver outcomes
Bill, retain, upsell, reinvest
Revenue from Layer E feeds back into Layer A — the loop never stops.
Three Loops That Make the OS Self-Sustaining
Discovery finds the battlefield. Delivery wins it. Revenue fuels the next cycle. Governance binds all three — weak evidence halts everything.
Observe → Hypothesize → Validate → Prioritize
Plan → Execute → QA → Operate
Bill → Retain → Upsell → Reinvest
Simple Outside.
Earning Universes Inside.
From the outside, MARIA OS presents a single clean interface. Inside, multiple autonomous “Earning Universes” orbit the core — each one a self-contained revenue engine.
Each Universe discovers its own opportunities and generates revenue independently.
The Agents That Make Laputa Earn
Each agent operates within a layer of the floating intelligence structure. Click a card for details.
Discovery Scout
Sky Sensor
Scans markets, news, and customer signals to surface opportunities.
Evidence Analyst
Evidence Fabric
Bundles and scores proof before any decision is made.
Decision Strategist
Decision Core
Evaluates ROI, risk, priority, and responsibility boundaries.
Execution Agent
Agent Exec
Deploys tasks, coordinates sub-agents, and delivers outcomes.
Revenue Optimizer
Revenue Loop
Bills, retains, upsells, and reinvests earned revenue.
Quality Auditor
Governance
Enforces governance gates and halts when evidence is weak.
6 Agents. 5 Layers. 1 Self-Earning Loop. — Click to inspect
Buy Skills. Bind to Agents. Start Earning.
Every skill is a revenue primitive. Browse, purchase, and deploy — agents consume skills to generate cash flow.
Market Signal Scanner
News, SNS, patent filings, regulation changes to surface opportunities.
Customer Intent Detector
Behavioral signals from CRM, web, and support tickets.
Evidence Bundler
Aggregates and scores multi-source proof before decisions.
Risk Scoring Engine
Real-time risk assessment with regulatory compliance checks.
Auto Outreach Agent
Autonomous email, call, and proposal generation for leads.
Contract Closer
End-to-end deal negotiation, pricing optimization, and signing.
Audit Trail Generator
Immutable decision logs with responsibility chain tracking.
Governance Gate Controller
HITL gates, escalation rules, and evidence threshold enforcement.
8
Skills Available
+310%
Avg ROI
3-20
Price Range (万/mo)
Agent Loadout
AGENT-PRIME
G1.U1.P1.Z1.A1
Total Stats
Total ROI
+0%
Monthly Cost
0万
Agent Profile
Purchase → Deploy → Earn → Reinvest
Skills are the unit of billing. Agents consume skills. Revenue funds more skills. The loop closes itself.
Browse
Marketplace
Scan skills ranked by ROI, category, and compatibility with your agent team.
Purchase
Billing
Subscribe per-skill with tier-based pricing. Pay only for what agents consume.
Bind
Deploy
Skills are dynamically bound to agents. Auto-refill keeps agents capable.
Earn
Revenue
Agents execute skills to generate revenue. Every action is metered.
Reinvest
Growth
Earnings fund new skill purchases. The loop compounds autonomously.
- Single agent binding
- Basic ROI tracking
- Community support
- Multi-agent binding
- ROI optimizer
- Priority support
- Auto-scaling
- Unlimited binding
- Custom skill creation
- Dedicated CSM
- SLA guarantee
- Outcome-based billing
Revenue = Skills Purchased × Agent Throughput × Conversion Rate − Skill Cost
Laputa Growth Hack Engine
Four autonomous squads form a self-reinforcing growth loop. They discover skills, deploy them, earn revenue, and reinvest — all without human intervention.
Discovery Squad
Find high-ROI skills before competitors
Continuously scans the Skill Marketplace for undervalued skills with high ROI potential. Buys early, deploys fast.
24
skills Found
+340%
avg R O I
78%
win Rate
Deployment Squad
Maximum skill utilization across agents
Optimizes skill-to-agent binding for maximum throughput. Monitors performance and auto-scales.
94%
utilization
0.3s
bind Time
99.9%
uptime
Revenue Squad
Convert skills into cash flow
Executes purchased skills against live opportunities. Maximizes conversion at every funnel stage.
480万
monthly Rev
156
deals Closed
2.1%
churn Rate
Growth Squad
Compound earnings into exponential growth
Reinvests earnings into new skills. Tracks portfolio ROI and prunes underperformers.
67%
reinvest Rate
+23%/mo
compound Growth
18
portfolio
The Laputa Flywheel
Discovery Squad finds skills → Deployment Squad binds them → Revenue Squad earns → Growth Squad reinvests
Every cycle compounds. The OS grows itself.
Watch Earning Universes Generate Revenue Autonomously
Each Universe processes tasks in real-time, earning revenue. They compete automatically — the top earner rises in the rankings.
Waiting for revenue events...
0
Transactions
—
Active Now
Sales
Leading
The OS earns. The OS evolves. The OS never sleeps.
Multi-Layer Economics
Three revenue layers compound as clients scale. The OS earns more as it delivers more value.
OS Subscription
Platform, governance engine, decision pipeline, audit trail.
10-200万/moAgent Deployment
Per-agent billing. Scale from 1 to full autonomous workforce.
5-20万/agentPerformance Fee
Revenue share on cost reduction and new revenue from agents.
10-30%100 clients × 3 tiers
~8億/year
OS 3億 + Agent 3億 + Outcome 2億
From Human Company to Agent Civilization
MARIA OS is not SaaS. It is an engine that builds companies. The evolution formula drives every cycle.
Xt+1 = E ∘ G ∘ J (Xt)
MARIA OS = Agentic Company Engine