A little tree that makes quick decisions.
A small bonsai with a private decision model inside. Ask it quick questions about text or photos, right on your home network. It answers with how sure it is.
Its canopy glows warm yellow whenever anyone uses it. That's all the light means.
In development. Not available to order.
Illustrative examples.
- Your shop's website asks: Is this review positive? Yes, 91% sure.
- Your inbox app asks: Which team should handle this email? Billing, 88% sure.
- Mom's phone asks: Which tags fit this photo of a dog on a beach? Dog, 96% sure. Beach, 93%. Snow, 2%.
- Your coding agent asks: How urgent is this, 1 to 4? 4 of 4, 84% sure.
It doesn't write. It decides.
A chatbot writes word by word. A decision model reads what you send once and picks from the answers you allow. Nothing gets written, so it's quick.
Every answer says how sure it is. It can be wrong, so that number matters.
One answer leads to the next.
Chain small answers and you get a decision tree. Here, one email finds its way in three questions. Thicker branches mean it's more sure.
Example numbers. One call can ask all three.
Your apps can ask it too.
Planned: a local API at decision-tree.local, behind an API key, with the same request shape as other decision-model APIs.
Example request · planned
POST http://decision-tree.local/v1/systemone
Authorization: Bearer <your key>
Content-Type: application/json
{
"state": "I was charged twice this month.",
"questions": {
"needs_reply": { "type": "noul", "instructions": "Does this email need a reply from a person?" },
"team": {
"type": "choice",
"instructions": "Which team should handle it?",
"criteria": { "billing": "payments, charges, refunds", "support": "using the product", "sales": "buying or upgrading" }
},
"urgency": {
"type": "score",
"instructions": "How urgent is this, 1 to 4?",
"criteria": ["can wait", "this week", "today", "right now"]
}
}
}Example response · illustrative numbers
{
"answers": {
"needs_reply": { "answer": "yes", "probabilities": { "yes": 0.93, "no": 0.07 } },
"team": {
"answer": "billing",
"probabilities": { "billing": 0.88, "support": 0.09, "sales": 0.03 },
"confidence": 0.84
},
"urgency": {
"answer": "today",
"probabilities": { "can wait": 0.04, "this week": 0.17, "today": 0.71, "right now": 0.08 },
"confidence": 0.66
}
}
}- Inside
- Planned: Raspberry Pi 5-class, low power.
- Power
- USB-C, a few watts (target). Solar kit planned.
- Reads
- Text and images. No camera or microphone.
- Model
- Planned: Liquid AI d1-omni-600M, open weights.
Want to know when it sprouts?
We'll write when there's real hardware, a price, and a date. We'd rather under-promise.
In development. Specifications, prices, and dates on this page are targets, not commitments, and may change before launch.