learn.agenticoutputs.com
Stop prompting.
Start building.
A learning portal for people building real things with AI — not just chatting with one. Start with the fundamentals if you need them, then spend five days going from your first agent to a deployed, observable system you'd trust in production. Practical, tool-by-tool, no filler.
Free, self-paced, and grounded in real tools — no signup wall, no filler modules.
Course Catalog
Choose your path
A fundamentals track to get you ready, and focused courses across agents, coding, orchestration, and retrieval.
Start Here
Practical Fundamentals
Linux and the terminal, Python environments, git, Postgres and Docker, and setting up an AI coding toolkit. No prior ML experience required.
Main Course
AI Agents & Vibe Coding
A 5-day arc from your first agent to a deployed, observable production system — tools, memory, skills, security, and shipping.
Optional
Local LLM Hosting
When and how to run a model on your own hardware with Ollama, instead of a hosted API — a short, focused add-on course.
Orchestration
Hermes Agent — From CLI to Production
Install, configure, and deploy the Hermes Agent platform end-to-end — memory, skills, gateway, cron, security, desktop plugins, and custom tools.
Coding Agent
Claude Code — From First Commit to Agent Teams
Anthropic's autonomous coding agent — print mode, interactive sessions, skills, hooks, subagents, MCP, CI/CD, dynamic workflows, and the Agent SDK.
Fundamentals
Prompt Engineering — From First Principles to Production
Design, test, and optimize prompts as an engineering discipline — zero-shot, chain-of-thought, structured output, RAG, safety guardrails, and prompt libraries at scale.
Data & Retrieval
RAG in Practice — From Naive Search to Production Retrieval
Build retrieval-augmented generation systems end-to-end — embeddings, chunking, hybrid search, re-ranking, agentic RAG, and production monitoring with pgvector.
Google Ecosystem
Google AI Coding Tools — Gemini CLI, Antigravity & Beyond
Google's agentic coding ecosystem — Gemini CLI, Antigravity IDE, Antigravity CLI, SDK, parallel agents, skills, browser-in-the-loop, and enterprise deployment.
Open Source
Terminal Agentic Coding Tools — OpenCode, Aider, OpenHands & Goose
Four open-source terminal agents compared — OpenCode (TUI with LSP), Aider (git-first pair programming), OpenHands (autonomous with Docker sandbox), and Goose (MCP-native, Linux Foundation).
The Flagship Course
Five days to a production agent
Hands-On Outcomes
What you'll ship
The ForemanCapstoneAn on-call assistant you page from Slack — researches, fixes, and remembers.
A Real Self-Improving Skill LoopDay 3Writes reusable SKILL.md files after solving a task — the actual Hermes Agent mechanism.
Stateful LangGraph WorkflowDay 3Multi-agent graph with checkpoint rollbacks and human-approval gates.
A Messaging Gateway SkillDay 5Expose an agent to Slack/Telegram via a real OpenClaw markdown skill.
Agent Memory PipelineDay 3pgvector storage, sensible chunking, and reranking for retrieval that's actually right.
Sandboxed ExecutorDay 2A locked-down Docker sandbox for running agent-generated code safely.
Observability StackDay 5Tracing, logging, and automated evals so you catch a regression before a user does.
Streaming Agent UIDay 5Real-time progress rendering via Server-Sent Events.Watch & Learn
Curated videos that pair with each phase of the curriculum.