Pre-launch · investor brief
A self-hosted operating system for human decisions.
AURA captures decisions and the reasons behind them on hardware the customer owns. Day-one ARPU is enterprise-tier. The moat is on-premise data sovereignty, not network effects.
The Problem
The why is never written down
Every AI assistant on the market follows procedures. None of them learns why a person chose what they chose, so none of them can act the way that person would. The decisions are visible. The reasoning dies with the person.
The Solution
A self-hosted Life OS
AURA captures the reason at the moment of the decision, structures it and cites it back, all on hardware the customer owns. No record, no answer. That neutralizes the largest objection to AI in high-trust settings.
Operators, principals, family offices
The Convergence
Why now
Meta validated the category
Meta Muse launched in September 2026 and became the most downloaded app in its first week. A personal agent with memory is now a mass-market want. Muse runs on Meta’s cloud. AURA runs on the customer’s hardware.
The generational transfer
Roughly $68 trillion in U.S. wealth moves to younger generations through 2045 (Cerulli Associates). The judgment behind that wealth is mostly undocumented. The assets transfer, the reasoning does not.
Local inference is ready
Open-weight models now run usable inference on a single server. The question has moved from “is it possible” to “who packages it for people who are not engineers.”
Annual contract value
Hardware-as-a-service tiers, billed monthly, plus a $5k one-time setup fee.
Hardware is included and passed through near cost. Software and service are the margin.
Revenue model
No ads, no data resale, no per-call billing. The customer owns the device. AURA earns on the service layer.
Legacy Vault and Editions add optional revenue after launch.
Why customers pay
The model and the data sit on hardware the customer owns. That property right is the whole value proposition.
Nothing leaves the premises. It keeps working with the internet off.
The moat
Big tech cannot copy this without breaking its own business model. Theirs is data harvesting. Ours is data ownership.
Cloud AI
- ✕Trains on user data
- ✕Memory held in the vendor’s cloud
- ✕Stops working if the vendor does
- ✕Renting intelligence
AURA Home Server
- ✓Never trains on your data
- ✓Memory on hardware you own
- ✓Keeps working without us
- ✓Owning your judgment
Actionable intelligence
Not a history book. A decision engine.
The question only your archive can answer
“Why did we pass on the Lisbon hotel deal?”
Result: AURA answers in the owner’s own words from March, cited to the call notes, the email and the voice memo. Fictional example from the product story.
Fifteen years later
“How did Mom decide when to sell a company?”
Result: A daughter with read access gets the pattern across six exits, grounded only in her mother’s records. Fictional example from the product story.
Why now, in three lines
Models that reason
Open models can now follow a person’s reasoning, not just their text. We can model the how, not only the what.
Agents that act
The market has accepted agents that work in the background. The open question is who the agent answers to.
Privacy as luxury
Trust in big tech is at a low. A server in your own home is the new premium.
The economics
Day-one ARPU is enterprise-tier because the value sits in an asset the customer ends up owning, not in metered usage.
Roadmap, after launch
Legacy Vault and Editions
Legacy Vault seals the customer’s Chronicle on their own hardware with copies in two other countries. The estate attorney holds the release. Each heir sees only what the owner chose for them. Memories marked private forever stay sealed, even from heirs.
Editions turn judgment into an asset while the owner is alive. The owner picks a part of the Chronicle, AURA removes other people’s private details and asks the owner to confirm, and the result is licensed for a term, revocable at any time. AURA takes a share of each license.
Where AURA stands today
Honest snapshot. No audited revenue and no public customer list yet. The bullets below describe the company as it is.
Investors
Request the brief
The full brief covers the product demo, the build status, the model and the plan for the first team.
