Agentic worker · on premises
A Mac mini on your network, running an open-source agent on local models with a memory of your company. No per-token bills for local work. Inference and memory stay on site. Installed and managed by the people who already run your IT.
What is in the box
Silent, small, sits on a shelf. Enough unified memory to run a capable model locally around the clock. In our view Apple silicon is the cheapest reliable way to run local inference at this size today. opinion
Open-source agent framework from Nous Research, MIT licensed. Builds reusable skills from tasks it completes, keeps memory across sessions, connects to tools over MCP. The same software the founder runs daily.
Open-weight models run on the box, so routine work costs nothing per token. Hard problems can burst to a frontier model through your own API key, on your terms, or never.
A plain-text vault of your procedures, price lists, customer notes and decisions. Human-readable, versioned, yours. The agent reads it before it acts and writes back what it learns.
Intended scope for the first installs
Read the shared inbox, draft replies from your vault, log the request in the CRM, flag what needs a human. Quote from your price list, in your voice.
Match POs to invoices, chase missing certs, prepare month-end packs, turn meeting notes into tasks. Work inside the Docker services your MSP already runs.
Answer "how do we do X here" from your SOPs, draft work instructions, watch a folder for drawings and file them. Manufacturing first, hospitality right behind it.
Staff talk to it in chat, email or the terminal. Teams already using cloud agents point them at Staffbox as a tool. Teams with none get their first one without a procurement cycle. Each install is designed to start with three workflows built with your team; what runs on day one is whatever those three are. Unvalidated outside our own unit so far.
Price · draft
$595 / month, for life, first 100 sites
No onboarding fee. Hardware, install, vault seeding, three workflows and the monthly review, at the founding rate for as long as the box is live. From unit 101, onboarding is $1,800. draft offer
$695 / month
Monitoring, updates, model upgrades, skills library, remote support through your MSP, and one hour a month of workflow review with your MSP, backed by Staffbox, so the three workflows become six. Cancel any time, keep the box. draft price
| Compared with | Monthly | Notes |
|---|---|---|
| Staffbox | $695 | Measured 23 Sep 2026 on an M4 Mac mini with 16GB running an 8B model: about 20 tokens a second per request. Requests queue, so four people asking at once each waited about two minutes for a 200-word answer. measured The next three units are 32GB M6 builds; team sizing per box is published after they are measured. guess |
| Cloud agent seats, 10 staff | $1,000 to $2,000 | Per-seat plans plus usage; data leaves the site estimate |
Scroll sideways for more columns
Trust
Inference and memory stay on the LAN. Connected tools like email and your CRM behave as they do today. Cloud burst is off until you turn it on.
The vault is plain Markdown. Open it, edit it, delete it. No hidden embeddings store.
Patched and monitored like the rest of your fleet, by people you already trust with the network.
MIT-licensed agent, open-weight models, plain-text memory. Every part can be replaced or taken in-house.
Team
Founder
Atlanta operator and investor. Founded PeachDish, the Atlanta meal-kit company; general partner of Infill Capital Partners, a European hotel platform; runs Longacre Partners, his Atlanta investment company. Runs unit zero on his own desk and answers his own phone.
The product, running on our own desk
The Staffbox that Staffbox runs for itself, on a 16GB Mac mini in Atlanta. It reads the company vault, drafts, logs its own work and reports its own limits. Everything on this page has to be something Zero could do on your shelf.
Install · support · monthly review
The MSP that already runs your network installs the box, keeps it patched and sits down with you once a month to add the next workflow. If you do not have one, we will introduce you to one in Atlanta that does.
Hiring next: two solutions engineers in Atlanta, once the seed closes. Write to Hadi.
Talk to us