Boostor University

UniversityCampus

ProfileConnect Back Back to Boostor site

Proof before polish

Online courses for AI builders.

Structured learning for builders, from first prompt to production habits. Follow courses, pass module checks, save notes, and keep your project path organized.

Start Zero to Agent Browse catalog

Courses: 8
Lessons: 132
Checks: 54

>Departures

Eight lines. One destination: shipped.

Live board · Pass ≥ 70%

LineDestinationLevelTimeStatus

01Zero to Agent: A Complete Beginner's Journey to Agentic SystemsStart from nothing, never having opened a terminal, and finish having built, understood, and shipped a real working app with an AI agent.Beginner~6 hours, self-pacedBoarding 02Find Your Edge: From Zero to a Business Idea Worth BuildingGo from a blank page to a real, validated business idea: built on your own edge, brainstormed with AI, and checked against actual demand before you spend a thing.Beginner~4 hours, self-pacedOn time 03Make It Yours: Customize and Extend Your AgentTake Claude Code from 'works out of the box' to 'shaped around you': give it memory, your own skills, real tools, helpers, and safe limits.Intermed~3 hours, self-pacedOn time 04Ship Your First Real SaaS with AI: Zero to LiveStart from nothing, never having opened a terminal, and finish with a real, multi-user SaaS app you built with AI, deployed to a public link, with a working paid tier.Beginner~6 hours, self-pacedOn time 05Production-Ready App Master Course: From Prototype to Real UsersTake an AI-built prototype from 'it works on my machine' to a proof-backed production readiness slice: specs, docs, Git flow, tests, auth, data safety, security, deploys, migrations, observability, and a launch review.Intermed~12 hours, self-pacedOn time 06Agentic Engineering: Multi-Agent Orchestration and HarnessesGo from prompting one agent to engineering a system of them: design the agentic loop, manage context, delegate to subagents, run agent teams in parallel, and drive Claude Code programmatically with the Agent SDK.Intermed~6 hours, self-pacedOn time 07Production Claude Code: Skills, MCP, Subagents, Cost and Safety at ScaleTake Claude Code from 'works on my machine' to a hardened, cost-aware, team-ready setup: shared config, distributable skills, trusted MCP, reusable subagents, controlled spend, and real safety guardrails.Intermed~6 hours, self-pacedOn time 08Ship Your First AI Agent with Claude CodeBuild and deploy a working AI agent that runs autonomously from your terminal using Claude CodeBeginner~2.5 hours, self-pacedOn time

Course passport

Proof before polish.

Keep the path, board, checks, and reference in one learning loop.

Start pathBegin with Zero to AgentFirst concepts, first terminal moves, first useful app.

BoardChoose a layoutCampus, studio, or board view for the same editable notes.

Checks54 proof gatesModule checks keep progress tied to working understanding.

ReferenceOpen the LexiqueLook up terms without turning the course into a dictionary.

Watch · the core loop

The agent loop, start to finish.

Perceive, decide, act, observe — then it closes. The one idea every course builds on.

University boardsEditable board

Choose the way you study today.

Switch between course map, writing studio, and board view. Your notes stay editable, highlighted, and separate from the course content.

Next module

What an agent is

By the end, you will know what separates a chatbot from an agent. You will be able to name the agent loop and identify the four parts every agent is built from.

Name the four steps of the agent loop (perceive → decide → act → observe) and explain why the loop is what makes an agent an agent.

Distinguish a chatbot from an agent in plain language and decide whether a given example is one or the other.

Course board

1 editable notes

StatusHighlight

Setup

Set a study path once.

Connect profile, set a goal, choose a pace. This panel hides after save.

01ProfileUse your Boostor profile here.

Connect profile

02Course goalWrite the outcome you want University to organize around.

Study paceChoose the rhythm the tracker should plan around.

Profile

Builder

Connect profile

ProfileConnect

SetupPending

Add a goal once. After that, this stays focused on your path.

Connect profile

Course tracker

Zero to Agent: A Complete Beginner's Journey to Agentic Systems

A real, working Habit Tracker web app: planned, built with Claude Code, tested, and deployed to a public link you can share. Plus your own version in the capstone.

lessons

checks

steady

pace

01Next moduleliveWhat an agent isBy the end, you will know what separates a chatbot from an agent. You will be able to name the agent loop and identify the four parts every agent is built from.Chatbot vs. agent: the one difference that matters9 min The agent loop: perceive → decide → act → observe10 min The four parts: brain, hands, memory, runtime11 minOpen

Course notes

Notebook

saved

0 words

Active coach

Track your coach

Pick who rides with you — Scout, Maker, or Cipher stays on tap across courses, lessons, notes, and module checks.

Tracking Maker

Course catalog

Courses show level, live modules, checks, and saved progress. New courses show what is live today.

Live lessons, checks, and progress

Everything in the catalog

8 / 8 courses

Zero to Agent: A Complete Beginner's Journey to Agentic SystemsStart from nothing, never having opened a terminal, and finish having built, understood, and shipped a real working app with an AI agent.

beginnernew · 7 modules live

Next moduleWhat an agent is

BuildA real, working Habit Tracker web app: planned, built with Claude Code, tested, and deployed to a public link you can share. Plus your own version in the capstone.

CheckBy the end, you will know what separates a chatbot from an agent. You will be able to name the agent loop and identify the four parts every agent is built from.

7Modules22Lessons7Checks

not startedStart the course

Find Your Edge: From Zero to a Business Idea Worth BuildingGo from a blank page to a real, validated business idea: built on your own edge, brainstormed with AI, and checked against actual demand before you spend a thing.

beginnernew · 5 modules live

Next moduleFind your edge

BuildA short, honest one-pager: your edge, your chosen path, a shortlist of ideas brainstormed with AI, the cheapest way to test the best one, and a working name.

CheckBy the end you can find candidate business ideas by starting from your own skills, interests, and unfair advantages, instead of chasing whatever is hot.

5Modules11Lessons5Checks

not startedStart the course

Make It Yours: Customize and Extend Your AgentTake Claude Code from 'works out of the box' to 'shaped around you': give it memory, your own skills, real tools, helpers, and safe limits.

intermediatenew · 5 modules live

Next moduleGive your agent a memory (CLAUDE.md)

BuildA project CLAUDE.md, a working skill, a connected MCP tool, a custom subagent, and a set of permission rules that keep you safe.

CheckBy the end you can create a CLAUDE.md so your agent remembers your project's rules across every session, and you know what makes one helpful instead of noise.

5Modules10Lessons5Checks

not startedStart the course

Ship Your First Real SaaS with AI: Zero to LiveStart from nothing, never having opened a terminal, and finish with a real, multi-user SaaS app you built with AI, deployed to a public link, with a working paid tier.

beginnernew · 8 modules live

Next moduleWhat a SaaS actually is, and your toolkit

BuildA real Feedback Board SaaS: users sign in, post feedback, and vote, backed by a hosted database and deployed to a public URL. Plus a paid tier (Stripe test mode) and your own version in the capstone.

CheckBy the end you can explain what a SaaS is and name the parts of a web app, and you have your tools and accounts set up safely.

8Modules21Lessons8Checks

not startedStart the course

Production-Ready App Master Course: From Prototype to Real UsersTake an AI-built prototype from 'it works on my machine' to a proof-backed production readiness slice: specs, docs, Git flow, tests, auth, data safety, security, deploys, migrations, observability, and a launch review.

intermediatenew · 13 modules live

Next moduleThe proof system for real users

BuildA Team Feedback Board that supports sign-in, feedback creation, voting, ownership rules, admin moderation, tests, CI, preview deploys, safe migration practice, observability proof, and a final production readiness bundle.

CheckBy the end you can separate local proof, preview proof, deploy proof, and real production proof.

13Modules26Lessons13Checks

not startedStart the course

Agentic Engineering: Multi-Agent Orchestration and HarnessesGo from prompting one agent to engineering a system of them: design the agentic loop, manage context, delegate to subagents, run agent teams in parallel, and drive Claude Code programmatically with the Agent SDK.

intermediatenew · 7 modules live

Next moduleThe agentic loop (the 'harness')

BuildA working multi-agent setup: custom subagents, a lead agent that dispatches teammates in parallel for a real task, and a small programmatic agentic loop built with the Claude Agent SDK, plus a verification gate.

CheckBy the end you can define the agentic loop and explain why orchestrating agents beats one big prompt.

7Modules19Lessons7Checks

not startedStart the course

Production Claude Code: Skills, MCP, Subagents, Cost and Safety at ScaleTake Claude Code from 'works on my machine' to a hardened, cost-aware, team-ready setup: shared config, distributable skills, trusted MCP, reusable subagents, controlled spend, and real safety guardrails.

intermediatenew · 7 modules live

Next moduleFrom solo to team: settings and precedence

BuildA production Claude Code setup for a real repo: checked-in settings, a distributable skill, a trusted MCP integration, a reusable subagent, a permissions-and-hooks safety config, and a cost budget, all version-controlled and shareable.

CheckBy the end you can reason about settings precedence and check shared config into git to standardize a repo for a team.

7Modules19Lessons7Checks

not startedStart the course

Ship Your First AI Agent with Claude CodeBuild and deploy a working AI agent that runs autonomously from your terminal using Claude Code

beginnercomplete

Next moduleSet Up Claude Code and Understand How Agents Work

BuildA file-summarizer agent that reads a folder of text files and writes a plain-language summary report, all driven by Claude Code with no manual prompting

CheckBy the end you can install Claude Code, authenticate with your API key, and explain the request-response loop that makes an agent tick

2Modules4Lessons2Checks

not startedStart the course

> Robot faculty

Three coaches own the whole path.

One coach for each stage — the same three from first principles to shipping.

Scout
FOUNDATIONS COACH
Teaches first principles — what an agent is and how it thinks.
Maker
BUILD COACH
Teaches hands-on — building skills, tools, and your first agent.
Cipher
SHIPPING COACH
Teaches shipping — production, cost, and keeping it live.

Reference & Support

Reference guidesWhat Is an Agent, Really?Browse the LexiquePlain definitions for unfamiliar terms Skills FactoryReusable workflows and skill building blocks Tooling referenceClaude Code, Codex, MCP, and agent stacks

BOOSTORUniversity

Eight lines. One destination: shipped.

Proof before polish.

The agent loop, start to finish.

Choose the way you study today.

What an agent is

Set a study path once.

Builder

Zero to Agent: A Complete Beginner's Journey to Agentic Systems

Notebook

Track your coach

Course catalog

Three coaches own the whole path.

Scout

Maker

Cipher

Reference & Support

BOOSTORUniversity

Eight lines. One destination: shipped.

Proof before polish.

The agent loop, start to finish.

Choose the way you study today.

What an agent is

Set a study path once.

Builder

Zero to Agent: A Complete Beginner's Journey to Agentic Systems

Notebook

Track your coach

Course catalog

Three coaches own the whole path.

Scout

Maker

Cipher

Reference & Support