Published
AI Agents Projects for Beginners
25 Projects to Build, Run, and Deploy Your Own Intelligent Agents
Twenty-five projects from hello-agent to a capstone deploy. Research, files, email, browser, multi-agent teams, and an observatory. Companion references for OpenClaw, Hermes, and Ollama. Build first. The model is a component, not the project.
Editorial principle
The student writes the agent. The book is a workshop, not a prompt pack. No hosted kids chatbot. Does not make AI safe.
- Chronology
- 2026
- Length
- ~119,000 words
- Reading time
- ~10 hours · 512 pages
Authors
Michael Gannotti
The voices
1 figures across 2026. Each speaks in their own voice, through their own texts.
The builder
nowProjects
Core question: What did you actually run?
The person who ships the agent — not the model that wrote a plan.
Chapters
25 chapters. The disagreement is not resolved.
- Complete
Project 1
Hello, Agent World
Can you run one agent end to end?
Voices: builder
- Complete
Project 2
Research Agent: Web Search Summarizer
Can it cite, or only compress?
Voices: builder
- Complete
Project 3
File Organizer Agent
What may it move, and what must it leave?
Voices: builder
- Complete
Project 4
Spreadsheet Intelligence Agent
Who owns the numbers after the agent runs?
Voices: builder
- Complete
Project 5
Smart Email Assistant
What must it never send?
Voices: builder
- Complete
Project 6
API Connector Agent: Data Pipeline
Where does the data go, and who approved it?
Voices: builder
- Complete
Project 7
Coding Agent: Helper and Refactor
Who reviews the diff?
Voices: builder
- Complete
Project 8
Personal Knowledge Base Agent
What stays local?
Voices: builder
- Complete
Project 9
Browser Automation Agent
When is a click a permission?
Voices: builder
- Complete
Project 10
Meeting Transcript Analyzer
Who is in the transcript, and who may see it?
Voices: builder
- Complete
Project 11
CI/CD Watchdog Agent
What fails closed?
Voices: builder
- Complete
Project 12
Document Generator Agent
Whose name goes on the page?
Voices: builder
- Complete
Project 13
Debate Team: Multi-Agent Consensus
Is consensus a vote, or a check?
Voices: builder
- Complete
Project 14
Code Review Panel
Who is accountable for the merge?
Voices: builder
- Complete
Project 15
Content Factory: Writer + Editor + Fact-Checker
What happens when the fact-checker loses?
Voices: builder
- Complete
Project 16
Customer Support Router
When must a human take the ticket?
Voices: builder
- Complete
Project 17
Data Quality Team
What is a silent bad row worth?
Voices: builder
- Complete
Project 18
Workflow Orchestrator
Who turns the pipeline off?
Voices: builder
- Complete
Project 19
The AI Analyst: Financial Report Reader
Can it be wrong with confidence?
Voices: builder
- Complete
Project 20
Smart Home Command Center
What must never fire unattended?
Voices: builder
- Complete
Project 21
Competitive Intelligence Monitor
What is public, and what is theft?
Voices: builder
- Complete
Project 22
The Ethical Guardian: AI Safety Auditor
Who audits the auditor?
Voices: builder
- Complete
Project 23
Deploy Your Agent: From Script to Service
What changes when it runs overnight?
Voices: builder
- Complete
Project 24
The Agent Observatory: Monitoring Dashboard
What do you measure if you cannot trust the logs?
Voices: builder
- Complete
Project 25
Your Capstone: Build & Deploy an End-to-End Agent
What did you refuse to automate?
Voices: builder