Drop in a skill. Your AI crew does the hard part.
VPS setup, server keys, the technical stuff that stops most people — your agents handle it. You answer one question and pick who does the work. No code. No fear. Built for people who've never touched a terminal.
Self-taught. Built it on the side. Not a coder by trade — but knows how to build a house from its foundation. That's all it took.
Drop it in. Answer one question. Walk away.
No setup ritual, no manual. Here's the whole thing, start to finish — the same way you'd explain it to a friend who's scared of computers.
Drop the skill in.
Drag a skill into the hub — like setting a book on the table. That's the whole start. No install, no config files, no command line.
Answer one question.
The crew asks: "First time, or technical?" You tap one. That tells it whether to just do the job for you, or walk you through it as it goes.
Pick a model, then walk away.
The architect sets up the files, wires the server and keys behind a safe, guarded flow, and stops to ask you about anything that matters. You pick the model and say go.
Built for the person who wants the result — not the syntax.
You don't need to know how to code. You need something you want to build — a website, a tool, a little business — and the willingness to drop a skill in and answer one question. The crew handles the technical part. You keep the money, and the credit.
See how you'll earn your first $50 →
Six small things nobody else is doing.
Type /goal. Walk away. Come back to ship.
Say what you want in plain language, pick which agents may run, set a budget and a stopping point — then close the laptop. The crew keeps going and only pings you when it actually needs you.
Approve a merge from your lock screen.
Hook up Telegram in one step. When the work's ready you get a ping — reply yes to ship it or no to hold. No laptop required.
Talk to the crew. Your voice stays put.
Tap the mic and just say what you want. It turns your voice into text right on your laptop — nothing leaves the machine. Pace the kitchen, dictate a goal, the crew picks it up.
Your second brain, on your disk.
Every goal, transcript, and summary is saved as plain notes on your own computer — yours to keep, search, and read forever. Open them in Obsidian and watch how it all connects.
The crew finds the right file the first time.
Instead of guessing, the crew builds an exact map of your project, so it lands on the right file every time — faster, cheaper, and without the fuzzy-search misses.
The system-prompt stopper.
Before your crew trusts a new skill or tool, a local scanner checks it for sketchy stuff — hidden instructions, leaked secrets, tampered downloads — and blocks the bad ones before they ever touch your machine.
Ollama runs the room from minute one.
You double-click the installer. Ollama is already on board — a real local model that routes, talks, and thinks for the crew. No accounts, no keys, no providers. You get a working AI before you've signed into anything.
- The brain is on your machine. Every word the crew says runs through local Ollama by default — your prompts and your code never leave the box.
- API rates? Not for you. Local routing means you don't burn provider tokens on the dispatch layer. The expensive calls are saved for the work that actually deserves them.
- Claude doesn't ship Ollama. RHOBEAR does. Other products gate everything behind a billing key. We gate nothing behind anything.
- Your monthly sub stretches further. When you do call Claude or ChatGPT, you're spending the flat sub you already pay — on the build for your business, not on the chitchat in between.
When you do want to plug in a paid model, here's why we don't nickel-and-dime you.
No meter. No bill that grows while you think.
Every other agent product asks for an API key. That key is a credit card the model bills every time it speaks — every token in, every token out. Three hours of conversation can cost more than a steak dinner. RHOBEAR doesn't take API keys. RHOBEAR signs in with the subscription you already pay for.
The meter never sleeps.
- Every word the crew says is metered against your card.
- One long conversation can ring up a steak-dinner bill.
- Forget you left a job running overnight? Wake up to a surprise.
- Switch models? New key, new dashboard, new way to leak it.
- Your Grok / ChatGPT Plus / Claude Pro sub covers the conversation. No extra meter.
- Use the same flat $20-ish monthly bill you've already paid.
- Tokens are not billed per-word. Pay-as-you-go goes out the window.
- No keys to leak. Revoke from your provider like any other app.
↳ Bring an API key if you want pure direct access — the option's there. But the default is the better deal: the sub you already pay for, doing more work for the same money.
It's easy.
The agent does the hard part. You drop in a skill, answer one question, and pick who does the work — no terminal, no keys, no setup ritual. Every other promise here is in service of that one.
Not just for engineers.
The skills catalog ships with vetted-safe modules across design, content, lead-gen, and developer work — with more verticals being added. Hire the crew once. Use them for every department.
/command Pops can run.
The hub. In your hands.
Monday.
We're shipping the public beta at noon Central · Monday, June 8, 2026. It's a small, limited beta — kept tight on purpose, so every run helps train the crew smarter and we can open it up to more people. Until then, here's the clock.
Buy once. Yours for life.
The beta is small and kept tight on purpose, and v1 lands shortly after. The first 1,000 hubs go out as Founder Edition — a one-time price, never a monthly bill, and free updates for life. Up to two hubs per address.
Prefer to boot from source while you wait? Here's the whole thing:
