BÁEZ LABS
Sovereign AI for Solopreneurs
A 2-day hands-on workshop where you build and deploy five real AI agents on your own machine — Chief of Staff, Content Agent, Finance Reporter, Follow-up Agent, Client Tracker. They run on a schedule. You get the output.
Freelancers, independent consultants, coaches, and one-person firms who want to stop paying $400–700/month in SaaS tools and start running the equivalent logic on their own machine — permanently, automatically, for the cost of an API call.
You don’t need to know how to program. You need to be comfortable in a terminal and willing to spend two days building something real.
8–10 participants per cohort, maximum. In-person (2 full days) or online (4 sessions of 3 hours each). Small by design — hardware setup at this level needs individual attention.
You leave with 5 agents running on your own Mac, a private GitHub repo with your code, and 30 days of async support.
What You Build in 2 Days
Session 1
Your Sovereign Stack + Chief of Staff Live (3h)
Why local AI vs. ChatGPT/Zapier: no monthly fee, your data stays on your machine. Install Ollama, pull qwen2.5:7b, wire up your Anthropic API key. Real agent architecture explained in depth — you're taking this home to maintain yourself.
Deployed live, already coded: Chief of Staff — reads your Notion, writes a weekly priority brief, sends it to your inbox every Monday, against your own real data.
Session 2
Automated Outreach + Content Agent Live (3h)
Gmail API + Google Chat webhooks, configured once. The plumbing that lets your agents send real emails and notifications without touching a third-party service.
Deployed live, already coded: Content Agent — turns your real rough notes into 3 LinkedIn drafts.
Session 3
The 3 Templates + Notion (3h)
Notion as your agent’s backend: read, write, filter — practical API in 20 minutes, not a course.
Delivered as templates, you deploy them with guidance: Finance Reporter (reads your bank CSV), Follow-up Agent (personalized prospect follow-ups), Client Tracker (active projects, last contact). These touch real numbers and unreviewed personalization — better deployed at your own pace than live in front of the group.
Session 4
Customize & Walk Away (3h)
pm2 + system cron: agents that start automatically on boot, run on schedule, restart themselves if they crash. Real time to tune all 5 to your own rules and prompts — unlike a team with IT behind them, you're the one maintaining this alone.
Close: every participant leaves with 5 agents running, a private repo, and a crontab that fires them on schedule. Stuck on one? We stay until it works.
What You Leave With
LIVE
Chief of Staff
Weekly priority brief: reads Notion + calendar, writes a plain-language summary of what to focus on. Lands in your inbox Monday morning. Runs on local model — zero API cost.
LIVE
Content Agent
Drop rough notes, get LinkedIn drafts back. Twice a month, automatically. Uses Anthropic API — ~$0.30/run.
TEMPLATE
Follow-up Agent
Drafts a personalized follow-up for every prospect you haven’t heard back from. You approve, you send. Uses Anthropic API — ~$0.20/run.
TEMPLATE
Finance Reporter
Weekly P&L from your bank CSV: income in, expenses out, what changed, what to watch. Runs on local model — zero API cost.
TEMPLATE
Client Tracker
One weekly report: active projects, last contact, next milestone, overdue items. Runs on local model — zero API cost.
What You Need Before You Arrive
This workshop runs on Apple Silicon only. The local models that make this work fit in 16GB of unified memory — no external GPU, no cloud rental, no $3,000 machine required.
Apple Silicon — M1, M2, M3, or M4 (any variant)
16GB unified minimum — 24GB+ recommended
40GB free
Updated to the latest stable release
Comfortable running commands — you don’t need to know Python, but you need to not panic at a terminal window
Not eligible: 8GB Macs, Intel Macs, Windows, or Linux. Spec verification required before your seat is confirmed. If your machine doesn’t qualify: full refund or hold on the next cohort.
Before Day 1: install Homebrew and Ollama (brew install ollama), pull a test model (ollama run qwen2.5:1.5b), and get an Anthropic API key with $5 in credits. That’s the full pre-flight.
Questions People Actually Ask
No. You need to be comfortable in a terminal, not write Python from scratch — you deploy 2 already-coded agents live and get the other 3 as templates to deploy with guidance.
It helps, but it’s not required — we can wire the agents to a fresh Notion workspace in the first session if you’re starting from zero.
Full refund, or we hold your seat for the next cohort — see Prerequisites above.
We stay until it works. Nobody leaves with a broken build.
Just your own Anthropic API usage — about $11/month for the two agents that call it. Everything else runs free, locally, forever.
Both formats run: 2 full in-person days, or 4 remote sessions of 3 hours each. Same content, same live builds, either way.
Not yet — today it summarizes your bank CSV (income, expenses, what changed). If 1099/quarterly-estimate prep is critical for you, tell us in the workshop and we can look at extending it on your own stack.
Founding Cohort — Launch Price
First two cohorts at founding pricing, per market. Regular pricing applies after that. Both markets launch simultaneously.
Includes: 2 days of instruction, private GitHub repo with your full agent stack, 30 days of async support. No upsells inside the workshop.
Reserve Your Seat in the Founding Cohort.
Minimum 8 people to run it. Once the group is full, the date gets locked in — your seat is held from the moment you reserve.