Kickoff
Build your AI agent and define the outcome.
This is not a library of videos. We build live. You leave with a deployed AI agent, one year of VPS hosting, six months of unlimited agent usage, a verified certificate — and, if you choose the business track, the complete system to find clients and close AI work.
A quick introduction to the live format, the projects and the outcome you leave with.
100+ AGENTS DEPLOYEDStefan Trbojevic is the founder of n8nlab, CEO of Taskwise.ai and Head of AI Automation across multiple online businesses. After deploying more than 100 agents and automation systems for companies of every size, he built VibeCode Academy to transfer the stack, judgment and delivery process that only real projects teach.
The founding cohort follows a fixed build sequence from use case to deployed project, commercial offer and final demo.
Build your AI agent and define the outcome.
Models, prompting, context and decisions.
RAG, documents, databases and state.
APIs, Gmail, Calendar and n8n.
Claude Code, Codex and websites.
Deployment, VPS and production operations.
Niche, offer, outreach and discovery.
Proposals, pricing, closing and final projects.
For one person, that means an agent handling leads, research, scheduling or operations while they sleep.
For another, it means the skillset, portfolio and sales process to turn AI implementation into income.
Do not choose by technical level. Both start from plain English. Choose by the life you want after the program.
Bring us a real workflow from your life or business. We help you turn it into an agent that understands context, uses your tools, asks for approval when needed and keeps running after the academy ends.
Build the agent foundation, then go further: Codex, Claude Code, websites, dashboards, client delivery, offers, prospecting, sales calls, proposals, pricing and closing.
01 export const agent = createAgent({ 02 role: "Revenue operator", 03 tools: [gmail, calendar, crm], 04 memory: longTermMemory, 05 approvals: "before-spend", 06 });
Every inclusion has one job: make sure the end of the cohort is the start of your leverage, not the end of your access.
We teach the full architecture so your agent can know things, remember, use tools, take actions, recover from errors and run safely in production.
OpenAI, Anthropic, Gemini, local models — the names will change. We teach how to decide based on context length, tool use, reasoning, latency, privacy and cost.
Explore the technical curriculum →From personal agents to multi-agent orchestration, these are the projects worth knowing. We teach the transferable architecture behind them — and how to choose the right tool for the job.
Personal agentHermes AgentOpen framework guide →
Personal agentOpenClawOpen framework guide →
Agent orchestrationLangGraphOpen framework guide →
Multi-agent systemsAutoGenOpen framework guide →
Agent teamsCrewAIOpen framework guide →
Data + agentsLlamaIndexOpen framework guide →
Enterprise agentsSemantic KernelOpen framework guide →
Typed agentsPydanticAIOpen framework guide →
Agent development kitGoogle ADKOpen framework guide →
Agent development kitOpenAI Agents SDKOpen framework guide →
Agent platformAgnoOpen framework guide →
Production AI agentsHaystackOpen framework guide →
TypeScript agentsMastraOpen framework guide →
Code-first agentssmolagentsOpen framework guide →
Autonomous agentsAutoGPTOpen framework guide →
Multi-agent softwareMetaGPTOpen framework guide →
Multi-agent researchCAMEL-AIOpen framework guide →
Stateful agentsLettaOpen framework guide →
Open agent frameworkBeeAIOpen framework guide →
Composable agentsAtomic AgentsOpen framework guide →
Model-driven agentsStrands AgentsOpen framework guide →
Software agent teamsChatDevOpen framework guide →
Software agentsOpenHandsOpen framework guide →
Enterprise orchestrationMicrosoft Agent FrameworkOpen framework guide →Project names and marks belong to their respective owners. Each card links to the official open-source project.
You learn to use Codex and Claude Code as build partners inside a repeatable loop: define, plan, build, verify, ship.
01 export const agent = createAgent({ 02 role: "Revenue operator", 03 tools: [gmail, calendar, crm], 04 memory: longTermMemory, 05 approvals: "before-spend", 06 });
Clients do not pay because you know Claude. They pay because you can identify an expensive problem, make the solution believable, remove risk and deliver an outcome.
The Business Builder track includes the complete client system.
See how we teach salesFour weeks. Live workshops. Build sessions. Office hours. Small cohort. You bring your project into the room and leave each week with something working.
Completion alone is not enough. Your final project is reviewed for function, integrations, safety, deployment and clarity. Pass the project and your credential gets a unique public verification ID.
This certifies that
has successfully designed, built, tested and deployed a functional AI system through the
The live cohort begins October 25. Once the remaining seats are gone, enrollment closes for this cohort.
* “Unlimited agent usage” applies to academy-managed agent usage for six months and is subject to fair-use, supported models and abuse protections. Details are listed on the offer terms page.
If you can commit four weeks and bring a real problem, we can start from almost any technical level.
Choose the path. Bring a real problem. We build from there.