The Problem
// why most AI tools fall shortYou've watched the agentic AI demos. You've tried the chatbots. They impress for 30 seconds, then forget who you are by Tuesday.
They forget
Every conversation starts from zero. No persistent memory of your projects, voice, or context.
They rent
Your data, prompts, and outputs flow to someone else's servers. Cancel the subscription, lose the work.
They sleep
They only work when you click. No autonomous execution, no cron jobs, no agent that wakes up at 2 AM and finishes the task.
They generalize
One-size-fits-all. They don't know YOUR workflow, YOUR voice, YOUR data, YOUR business.
The Promise
// what you'll have when you finishYou won't just understand agentic AI. You'll own one — running on your hardware, knowing your context, executing while you sleep.
Most AI courses teach concepts. This one ships a system. By the final module, you'll have a working multi-agent setup customized to your use case — content creation, trading, research, solopreneur ops, or whatever you bring — with persistent memory, scheduled execution, and the ability to spawn subagents on its own.
It's the same architecture I use to run Hermes, my personal multi-agent system that operates 24/7 across my home GPU and a VPS, syncing through an Obsidian vault and reporting back when it's done. You'll build the simplified version, then customize it to your life.
What You'll Build
// the full architectureEvery layer is taught, demoed, and built from scratch in the course. No black boxes.
Who This Is For
// honest fit-checkThis is for you if…
- You want leverage from AI without wiring together 14 SaaS tools
- You're comfortable in a terminal (or willing to be)
- You want an agent that knows YOUR context, not a generic chatbot
- You're tired of toy demos and want something running in production
- You believe local-first compute matters (privacy, cost, speed)
- You want to stop renting AI and start owning it
Skip this if…
- You want a no-code "click to deploy" AI assistant — this teaches you to build
- You're looking for ML theory or research papers
- You expect to ship in 1 hour with no setup
- You're not willing to spend ~$5/mo on a VPS for the always-on parts
- You don't want any direct contact with code
Pick Your Path
// Module 7 — customizing Hermes for YOUIn Module 7, you customize Hermes to your specific life. Pick one path or combine them. The framework adapts.
Hermes drafts posts in your voice, schedules across platforms, analyzes what works, refills the queue while you record.
Watches markets, ingests news, runs strategies on a schedule, alerts you only when something matters.
Triage inbox, qualify leads, draft replies in your tone, route customer support, generate reports each Monday.
Monitors arXiv + RSS, summarizes papers, connects ideas to your existing notes, weekly digest in your inbox.
Bring your own use case. The Hermes framework adapts to anything where memory, autonomy, and personalization matter.
Curriculum
// 10 modules · video + code + hands-on00 Welcome & Setup 30 min
What is agentic AI (and what it isn't). The Hermes philosophy: local-first, autonomous, persistent. Tour of the finished system you'll build.
Hands-on: Install Claude Code, verify your first agent run.
01 Foundations: Agents, Tools, and Context 1.5 hr
The mental model: LLM + tools + memory + loop = agent. Token economics. Why agents forget, and what to do about it.
Hands-on: Build a "hello world" agent that reads, summarizes, and saves a file.
02 Claude Code Mastery 2 hr
Skills, hooks, slash commands. MCP servers. settings.json deep dive. Writing your first custom skill and your first hook.
Hands-on: Build a personal "code reviewer" skill that runs on every commit.
03 Local-First Compute 1.5 hr
LM Studio vs. Ollama vs. cloud APIs. Model routing: Haiku cheap, Sonnet hard, Opus reasoning. Securing a local LLM with bearer tokens.
Hands-on: Run a model locally and route an agent to it.
04 Persistent Memory & The Vault Pattern 1.5 hr
The Obsidian vault as agent memory. Inbox/outbox protocol. Identity files (SOUL.md, MEMORY.md, BOARD.md). Auto-sync without conflicts.
Hands-on: Set up your vault, write your SOUL.md, sync via git.
05 Going Always-On: VPS & Cron 2 hr
Why $5/mo is enough. SSH, Tailscale, zero-trust networking. The brain-cycle pattern: VPS asks local LLM, executes, syncs back.
Hands-on: Deploy a VPS, install your agent, schedule nightly runs.
06 Multi-Agent Orchestration 2 hr
Specialist agents: planner, executor, reviewer. Brief/ack protocol. Parallel vs. sequential. Avoiding context contamination.
Hands-on: Build a 3-agent system: research → plan → execute.
07 Customizing Hermes for YOUR Use Case 2 hr
Pick your path: Content creator · Trader/analyst · Solopreneur · Researcher · Custom. We adapt the framework to YOU.
Hands-on: Build the version of Hermes specific to your life.
08 Productionization 1.5 hr
Secrets & rotation. Logging & alerting. Cost monitoring (no surprise bills). Failure modes and graceful degradation.
Hands-on: Add monitoring, alerts, and a kill switch.
09 Going Further (+ Capstone) 1 hr
Voice (Whisper + TTS). Vision (screenshots, OCR). The web (browser automation). Live walkthrough of my actual Hermes setup.
Capstone: Ship your customized Hermes. Submit a 5-min walkthrough. Featured students get a stream shoutout.
Pricing
// founder pricing for first cohort — May 4, 2026Get started today
- "Your First Agent in 30 Min" guide
- Module 0 video preview
- Public Discord access
- Stream notifications
Coffee-money entry
- Modules 0–2 (videos + code)
- Lifetime access & updates
#foundationsDiscord channel- Test the course at low risk
Most accessible full course
- All 10 modules (video + code)
- Lifetime access & updates
- Full student Discord
- Async capstone review (written)
Live + community · starts May 4
- Everything in Self-Paced
- 6-week live cohort with weekly streams
- Group office hours, 2×/week
- Cohort Discord with peer accountability
- Live Q&A during streams
Personal guidance · 5 seats
- Everything in Cohort
- 4 × 60-min private sessions
- Custom architecture review
- Priority Discord support
- Direct access for 90 days
Capped: 2-3 students total
- Everything in 1:1
- I build YOUR Hermes WITH you, live
- 8 × 60-min sessions over 8 weeks
- Custom skills & hooks built together
- 6 months priority support
Add-ons
Book any time. 60 minutes. Bring your stack, leave with a plan.
Send me your setup. I review it. You get a written report with prioritized fixes.
Reach builders and engineers who care about agentic AI. For relevant tools and platforms.
Who's Teaching
// not a guru, a practitionerI'm not selling theory. I'm selling the system I run every day.
I'm Alex Neff — creative technologist and builder. I shipped MacroBox, a Windows automation tool that delivered 3-7× throughput on Rivian's autonomy annotation pipeline at 99.9% accuracy, adopted by 15+ teammates with training I authored solo.
Outside that, I built and operate Hermes: a local-first multi-agent system running 24/7 across a home GPU and a VPS, coordinating through an Obsidian vault, executing scheduled tasks, and reporting back. It's the system this course teaches you to build, in simplified form.
I stream the work weekly on YouTube, Twitch, and X. The course is the structured path that takes you from zero to your own version of what you see on those streams.
First Cohort Schedule
// May 4 → June 14, 2026 · 6 weeksLive streams: Tuesdays + Thursdays · Office hours: Wednesdays + Saturdays · All sessions recorded.
Free Starter Guide
// 30 min · zero costGet "Your First Agent in 30 Minutes"
A focused walkthrough that takes you from zero to a working agent in half an hour. Read it free, no signup needed. Drop your email if you want stream and cohort notifications.
Read the guide now →FAQ
// the real questionsDo I need to know how to code?
Comfortable in a terminal helps. We'll teach the basics if not. You don't need to be an engineer — knowledge workers and builders ship great Hermes setups in this course.
What hardware do I need?
Any laptop works for the course. Modules 3+ benefit from a GPU for local LLMs (anything modern works). Module 5 introduces a $5/mo VPS for always-on capability.
Will this work on Mac / Linux / Windows?
Yes — all three. I run Hermes primarily on Windows + a Linux VPS. We cover platform differences in each module.
How is this different from a "ChatGPT for X" course?
Those teach you to use someone else's agent. This teaches you to build your own — owning the data, controlling the cost, and customizing every behavior. You're not renting; you're building.
What if I get stuck?
Foundations and Self-Paced students get the Discord and async capstone review. Cohort students get live office hours. 1:1 and Done-With-You students get direct access to me. Pick the support level that matches how stuck you expect to get.
Can I get a refund?
14-day refund window on all tiers, no questions asked, as long as you've completed less than 50% of the course (we can see this).
When does the next cohort start?
First cohort: May 4, 2026. Subsequent cohorts run quarterly. Self-Paced and Foundations are always available.
Will this stay current as AI changes?
Yes. Lifetime access includes all updates. The course architecture is deliberately model-agnostic — when Claude 5 ships, you swap a config line.
Can I upgrade my tier later?
Yes. Email me, you pay the difference, you're in. Foundations → Self-Paced → Cohort → 1:1 → Done-With-You — upgrade any time.
Is the content recorded? Can I rewatch?
All cohort streams are recorded. You keep access to videos, code, and exercises forever. Self-Paced students get the same library, just without the live calendar.
Ready to stop watching and start building?
Founder pricing closes when the first cohort fills. Self-Paced and Foundations are always open. Free starter is, well, free.