Build Intelligent AI Systems That Work While You Think
Welcome to AI Agent Systems, where you learn to build autonomous agents and intelligent workflows that scale your expertise without scaling your workload.
If you’re ready to go beyond automation and create AI systems that think, decide, and execute independently, you’re in the right place.
What You’ll Find Here:
Complete implementation guides for building autonomous research agents, content generation systems, and knowledge retrieval agents from scratch
Production-ready code examples and system architectures for RAG, MCP, and multi-agent orchestration
Real case studies from shipped applications: note generators, research systems, and data-driven tools serving real users
Step-by-step frameworks for turning your domain expertise into AI agents that work independently while you focus on strategy
Hi, I’m Jenny 👋.
I build AI systems and write about what happens when they run. Not what’s possible. What actually works, and what breaks.
This is the Agents Hub, other hubs you might also enjoy:
Claude Hub · Vibe Coding · AI Agents · Shipped Products · Substack Growth
Understand AI Agents
🤖 AI Agents and Automation: Everything I’ve Built, Tested, and Run in Production
Everything I’ve built, tested, and run in production around AI agents and automation
A single map of patterns, tools, and lessons across years of shipping
🔁 Claude Code Routines Are Hugely Underrated. Here’s How to Get Ahead of 99% of Users
The routines that act like always-on agents inside Claude Code
Setup patterns for daily, weekly, and event-driven jobs [PAID]
🧠 AI Agents for Everyone: 3 Types You Already Use Without Knowing It
What actually makes something an “agent” vs. a chatbot
The universal pattern behind every AI agent system
How to think about agents before building one
🔄 Agentic Flywheels: When AI Products Start Running (and Growing) Themselves
What happens when agents compound
The feedback loops that make AI systems self-improving
🤝 When AI Becomes Your Partner, Managing It Becomes Your Job
The management layer nobody talks about
How to supervise AI systems without hovering
Build Your First Agent
📝 Build a Content Generation Agent
Generate 18 variations per article automatically
Deploy to production and serve real users
Handle async processing and credit management
📧 Build an Email Processing Agent
Connect to Gmail API and extract content automatically
Build scoring systems for content quality and relevance
Generate summaries and prioritize reading
📊 Build Data-Driven Applications
Design database-backed analytics systems
Build real-time dashboards with automated updates
Deploy with Next.js and Vercel for scalability
🎭 Orchestrate Multiple AI Agents
Cast different models for different roles
Build multi-step research workflows with checkpoints
Avoid the AI Nanny Syndrome with structured supervision
Advanced Agent Systems
🔬 Build an Autonomous Research Agent
Set up autonomous pharmaceutical research system in 15 minutes
Adapt the framework to market research or competitive analysis
Build 3-phase methodology for any research domain
🧠 Build Your Second Brain with RAG
Set up semantic search across your content
Create chunking strategies that preserve meaning
Build chat interfaces with source citations
🔗 Connect AI to Your Data with MCP
Create your first MCP for database queries
Connect Claude Desktop to your data sources
Eliminate manual SQL and data retrieval friction
🛠️ Every Way to Build an MCP Server (From the Ones I Actually Shipped)
Complete guide to stdio vs HTTP MCP servers with 5 distribution patterns
Step-by-step builds for local, npm, .mcpb, Bearer token, and OAuth PKCE
Starter templates and real examples from 6 shipped MCPs [PAID]
🛠️ The Custom MCPs Included in Your Build to Launch Subscription
Overview of the research, writing, Substack, and workflow MCPs available to paid members
Build to Launch MCP, Substack Notes MCP, Substack Article MCP, and more
How to access and start using them
🛠️ Best MCP Servers for Claude Code: Replace Your Workflow, Keep Your Brain
The MCP servers I actually use daily
Research, sell, build, publish, debug — without breaking your chain of thought
📊 I Built an AI Research Tool to Study Substack Creators I Admire. Here’s What 3,000 Notes Show.
Build a custom MCP server to study engagement patterns across 9 creators
Discover 5 universal patterns from 3,230 notes analyzed
Learn the three-pass research methodology: classify, analyze, find contradictions
🔍 Perplexity + Claude in Chrome: How I Built a Competitor Research Workflow in 2 Hours
Connect Claude Code to Perplexity via a Chrome MCP bridge
Competitor research, market analysis, and content research without switching tabs
Built and deployed in 2 hours [PAID]
🚀 Claude Dispatch: The 80% the Demos Don’t Show
18 tasks tested across all 5 Claude remote modes: Dispatch, Scheduled Tasks, Computer Use, Chrome, code triggers
Exact prompts, real times, honest verdicts on what works and what fails
The prompt structure that prevents most stalls and the stack that makes it all possible [PAID]
🗓️ Claude Cowork vs Local Routines vs /loop vs Cloud Routines: What’s Actually Different
Four Claude scheduling modes mapped: /loop, Cowork Scheduled Tasks, Local Routines, Cloud Routines
The one failure-cost question that routes every automation job to the right tool
When OpenClaw and n8n still win — and why 80% of automation jobs don’t need them
🤖 Notion AI Agents: Real Examples, Limitations, and How to Build Custom Ones
Real examples of AI agents built on Notion
How to integrate AI workflows with your existing workspace
🎙️ When Typing Became the Bottleneck: How to Build a Voice-First AI System
When keyboard input becomes the constraint, switch to voice
Build voice-first AI into your workflow
🔧 Every Way to Build an MCP Server (From the Ones I Actually Shipped)
Complete guide to building MCP servers: HTTP, stdio, and hybrid approaches
Real examples from shipped MCP servers with production code
Deployment, testing, and maintenance best practices [PAID]
🔧 The Custom MCPs Included in Your Build to Launch Subscription
Overview of custom MCP servers available to Build to Launch subscribers
How to access and use them in your Claude Code and Cursor workflows
Examples of real workflows powered by these MCPs
🔬 How to Do Research With AI Effectively: 3 Questions Walked Through Live
Live walkthrough of 3 research questions using AI tools
Framework for structuring AI-assisted research projects
How to validate findings and avoid AI hallucinations [PAID]
Agent Platforms
🐙 How to Install OpenClaw and Run Your First Autonomous Agent
Local and Oracle ARM install side by side
Identity files (SOUL.md, AGENTS.md) that constrain the agent safely
Telegram pairing + first cron job firing without being asked [PAID]
🔀 4 Levels of AI Automation: When Claude, n8n, and OpenClaw Each Win
The four automation levels: output, workflow, decision-making, self-improvement
When to use Claude/ChatGPT vs. n8n vs. OpenClaw agents
The decision rule: “How much of the system does AI need to own?”
🐙 What Happened When I Let OpenClaw Run My Business for Four Weeks
Office hour recording: 18 cron jobs, Telegram interface, autonomous social media evolution, SEO workflow
The flip-flop moment: when workflow tools hit the judgment layer
Three-phase social media autonomy: distribution → human-in-the-loop → full autonomy
🐙 OpenClaw for One-Person Businesses: What It Is, What I Tried, and What Matters for You
The full picture of OpenClaw after weeks of running it
Framework for deciding if it’s for you
See Agents in the Wild
Real builders shipping real agent-powered products:
He Built an MCP Server So You Can Talk to AI About Any Substack Author -- Alejandro’s Substack Author MCP
Your Memories, Your Terms: The AI Memory Layer That Keeps Users in Control -- building AI memory with user ownership
Frequently Asked Questions
What’s the difference between automation and AI agent systems?
Automation executes predefined sequences. AI agent systems think, decide, and adapt independently based on your goals.
Do I need to be a developer to build these systems?
You need curiosity and willingness to learn. I didn’t have a CS degree when I started.
What tools and technologies do you use?
Python for agents, Next.js for web apps, RAG with embeddings, MCP for data connections, Cursor for AI-assisted development.
How long does it take to build an autonomous agent?
First research agent: 15 minutes. Production content generation app: 2-3 weeks.
Can these systems work with my specific domain or data?
Yes. The frameworks are domain-agnostic.
Start Building Your First Agent
👉 New to AI agents? Start with AI Agents for Everyone to understand the pattern
👉 Want to build research systems? Jump into Domain-Specific AI Research Agents
👉 Ready to ship production apps? Follow the journey from prototype to production
👉 Building something cool? Share it on Vibe Coding Builders
Browse all articles: AI Agents · Generative AI Projects
Updated May 04, 2026