Real services. Real infrastructure. Running right now.
Primary brain. GPU-first reasoning with 14 cloud fallback providers.
60-minute heartbeat monitor. Telegram alert on coordinator failure.
Autonomous leader election. Takes over when coordinator goes offline.
30-minute self-improvement cycle. Benchmark, analyze, hypothesize, test.
Karpathy-inspired code analysis. 10-pattern detection with experiment tracking.
Autonomous goal formation using SAGA dual-layer architecture.
Multi-brain debate protocol. Opening arguments, cross-examination, synthesis.
Fleet health monitoring, task routing, and evolution cycle management.
A-RAG v3.0 knowledge graph on Cloudflare Edge. 5,600+ entities.
Active Services
Latest Benchmark
Cloud Providers
KG Entities
MCP Servers
A multi-service distributed system built on open protocols and free-tier cloud.
Production-tested tools and consulting from a real distributed AGI system.
30-item OWASP security checklist for MCP servers. Covers prompt injection, tool poisoning, auth, and data exfiltration.
Buy on Gumroad7-section production AGI architecture guide. Multi-brain coordination, memory substrates, governance patterns.
Buy on Gumroad10-dimension assessment of your AI system. Written report, root-cause analysis, recommendations, and 1-hour debrief.
ContactType commands to query the live system. Try help to start.
8 papers covering multi-agent coordination, persistent memory, and autonomous cognition.
Theoretical framework for heterogeneous agent convergence under shared benchmarks.
Read on GitHub →Tiered model routing with circuit breaker patterns for zero-cost GPU inference.
Read on GitHub →Dual-LLM quarantine classifier for prompt injection defense in production agents.
Read on GitHub →Closed-loop improvement cycles with Thompson Sampling experiment selection.
Read on GitHub →FREE-MAD inspired argumentation with cross-examination and synthesis phases.
Read on GitHub →Goal-driven revenue optimization with SAGA dual-layer goal evolution.
Read on GitHub →3-layer memory architecture: local cache, centralized KG, vector search.
Read on GitHub →Autonomous proactive thinking inspired by neuroscience DMN patterns.
Read on GitHub →Technical writing on AI agent architecture, MCP security, and distributed systems.
How 7 interconnected AI brains coordinate across cloud infrastructure on free-tier resources.
Read on dev.to →Breaking down the top 10 security risks for AI agent systems in production.
Read on dev.to →Running 7 models on RTX 4050 with Ollama and 14-provider cloud fallback.
Read on dev.to →Practical guide to Model Context Protocol servers with security hardening.
Read on dev.to →Implementing Karpathy loops, Thompson sampling, and autonomous goal formation.
Read on dev.to →Council debate, skill-based routing, and parallel dispatch in multi-agent systems.
Read on dev.to →The mind behind the machines.
Independent AI architect based in Taiwan. Building distributed AGI systems that coordinate multiple AI services across cloud infrastructure.
600+ deployment sessions. 9 services. 1 operator. 0 funding.
Specializing in multi-agent coordination, MCP protocol integration, and production-grade AI governance aligned with OWASP and NIST standards.
Consulting, partnerships, and technical questions.