DBATR · Financials as a whole · Research & development as a whole
One ledger. Open on both sides.
Money is a tool for reaching people, not a scoreboard. Every dollar in has a named source and a nameable return. Every dollar out lands on a family, a build, or the people doing the work. Research and development sits inside that ledger as a protected floor — because invention is the only reason the core can stay free for the person who needs it most.
Free to the person. Priced to the organization. Published either way.
Money in
Shares are of the operating year, not of a fundraising target. No line here charges a person seeking help.
Licensing — patents and platform
40%Health systems, agencies, insurers, and operators license the filed inventions plus the coordination platform annually. Highest value per contract, lowest cost to serve.
Annual license, five figures and up, per organization
Facility and provider partnerships
18%Tiered flat monthly fee by licensed bed count. No paid placement, no pay-for-ranking, ever.
$79 / $149 / $249 per month
Coordination engine, metered
14%Base fee plus a charge per coordinated case actually resolved. Nobody pays for a case that went nowhere.
Base + per resolved case
Employers and workforce programs
10%Per covered employee, per month. Confidential routing that never reports back to the employer.
Per covered employee, per month
Government program contracts
9%Scoped municipal and state contracts with public reporting against agreed measures.
Negotiated program scope
Training and professional education
5%Per seat or per cohort, with a real credential recorded in the public verification registry.
Per seat / per cohort
API and white-label access
4%Other software companies build on the matching and coordination endpoints. Free trial tier, then metered.
Metered calls and matches served
Money out
In order of claim. Wages are funded first, R&D second, and administration is deliberately small and deliberately real.
The people doing the work
46%Wages on published bands, benefits from the floor up, and paid apprenticeships. This line is protected before anything else is funded.
Published pay bands, no salary history inputs
Research and development
22%Engineering, model work, field trials, and the patent pipeline. Held as a fixed floor, not a leftover — it is what keeps the free core free next year.
Fixed floor of operating revenue
Direct help to families
17%Rides, deposits, vet bills, gear, household goods, filing costs, and the pieces that unblock a person the same day.
Cost of removing a named barrier
Infrastructure and delivery
9%Hosting, phone lines, compute for on-device and hosted models, and the old-device support that keeps it usable.
Metered platform cost
Administration, legal, audit
6%Bookkeeping separated from the for-profit entity, counsel, insurance, and outside audit. Deliberately small and deliberately real.
Actual cost, published annually
The rules that govern the money
Never charge a person in crisis
The core stays free forever for individuals and families: matching, crisis routing, peer support, resource lists, reporting a facility.
Why: A price between a person and care is the failure we exist to fix.
Every payer receives something nameable
If we cannot state in one sentence what the payer gets back, we decline the money.
Why: That single test rules out kickbacks, paid placement, and hardship appeals at once.
R&D is a floor, not a remainder
Research and development is funded before growth spend, and its share is published.
Why: The free core survives on invention. Cut R&D and the whole model quietly becomes charity.
One ledger, open
Sources, shares, and outcomes are published on a fixed schedule, including the years that look bad.
Why: Trust is the only asset that cannot be licensed or replaced.
No funding with conditions on who we serve
Money that narrows who gets help is returned.
Why: Needs of the many. That is the whole point.
Reserves before expansion
Hold operating reserve before adding a new region, program, or headcount class.
Why: A program that collapses mid-year does more harm than one that never opened.
Research and development, as a whole
Staged by how close each line is to a person's hands. Nothing here is a demo — every line has a family it is meant to reach.
in the field
Coordination engine
Problem: One family retells the same story to nine agencies, and nobody owns the outcome.
Work: One intake that routes across care, legal, housing, transport, and work, with a consent record and audit trail per case.
Payoff: Resolved cases instead of referrals. It is also the metered revenue line.
in the field
Care matching without paid placement
Problem: Bed search is a phone tree, and ranked lists are usually bought.
Work: Live capacity, insurance, and fit matching, scored on the person's needs with reasons shown.
Payoff: Faster placement and a search result nobody can purchase.
in build
On-device assistance for old hardware
Problem: The people who need help most have the worst devices and the least bandwidth.
Work: Small models running locally with offline-first flows, so tutoring, translation, and paperwork help work with no signal.
Payoff: Privacy by default and a working tool on a five-year-old phone.
in build
Filing and packet generation
Problem: Procedure, not merit, decides most family and housing outcomes.
Work: Deadline extraction from notices, motion drafting, exhibit assembly, and a human signature on everything that matters.
Payoff: A self-represented parent files on time and on the record.
in build
Schedule-aware orchestration
Problem: Nobody can juggle work, court, treatment, school, and care on their own calendar.
Work: An assistant that plans around each person's real constraints — shifts, custody days, transit, sobriety schedule — and moves the pieces itself.
Payoff: The plan bends to the person instead of the person bending to the system.
in research
Early signals, handled as care
Problem: Hunger, housing loss, and crisis get noticed by discipline systems first.
Work: Signal detection that surfaces to counselors and advocates only, with no law-enforcement pipe and no surveillance product.
Payoff: Help arrives before punishment does.
in research
Live two-way language
Problem: Families are shut out of their own meetings by language.
Work: Low-latency two-way translation for conferences, intakes, and hearings, on the family's device.
Payoff: A parent participates instead of being described.
How we decide what to build
- 1Does it reach a person who is currently unreachable?
- 2Does it work offline, on an old device, in a bad month?
- 3Can a human check the output before it affects somebody's life?
- 4Does it survive us — documented publicly so another company can build it too?
- 5Does it lower the cost of the free core rather than raise it?
If a line fails one of those, it does not get funded — however impressive it looks. Want the specifications? They are published on capabilities, free, so another company can build it too.