You're pledging to donate if the project hits its minimum goal and gets approved. If not, your funds will be returned.
Hospitals, county emergency offices and water utilities now coordinate through a few central cloud services. When those fail (a storm, a grid outage, a cyberattack, or one bad software update like July 19, 2024, when about 8.5 million Windows machines crashed and some 911 lines went down), people fall back to paper, radios and phone trees.
I'm building ArkNet: software that keeps shared status, tasks and records working locally across devices a site already owns when the center is gone, then re-syncs when it returns. AI can help, but no automated action happens without an accountable human.
Already live (built solo): Aegis, a human-approval checkpoint for AI agents (aegis-mauve-one.vercel.app; sandbox plus enforcement API, not yet production security). Continuity Audit, which scores whether an organization can actually recover (arknet-continuity-audit.vercel.app; working MVP). CityOS, a flood-risk console for city operators (cityos-nu.vercel.app; prototype on sample data).
Goal: find out, with real operators, whether offline-first coordination is something they will drill on and pay for, before building more.
Days 0-30: publish the reference architecture and run 20+ interviews with emergency managers, hospital IT staff and utility operators, starting in Georgia.
Days 31-60: a working demo in which three ordinary devices keep sharing status and tasks with the internet cut, then reconcile cleanly when it comes back.
Days 61-90: one design partner (ideally a county emergency office or small utility) runs a tabletop outage drill on the demo; an independent engineer reviews the architecture.
Kill signal: if operators won't commit to a drill after seeing the demo, I narrow the use case before spending more. I'll post updates here at each milestone.
Minimum $5,000: devices for the outage demo plus an independent architecture review. I build the demo myself (slower, about 4 months).
Goal $40,000: $24,000 contract distributed-systems engineer (offline-first/mesh) for 3 months, my first hire; $6,000 devices and field equipment; $4,000 cloud and AI costs for running the pilots with real users; $3,000 independent technical review; $3,000 US incorporation, pilot-agreement legal terms and domestic travel to the design partner. No salary for me.
Just me for now: Michael O. Ogunleye (linkedin.com/in/michael-o-ogunleye-997938371). Since 2023 I've designed, coded and shipped every ArkNet product myself with no outside funding: the three pilots above plus three live consumer AI products (aihuman.digital, ainegotiator.digital, aicertified.digital). Since 2021 I've also been Chief Strategy & Technology Officer at Golden Royal Intercontinental. I'm based in the Atlanta area; ArkNet's business name is registered in Nigeria (OPULENTIA DIGITAL CORE HORIZON, RC 9802444).
Most likely causes: (1) operators like the idea but won't commit time to a drill, because public agencies move slowly and trust established vendors; (2) the offline sync turns out harder than planned on mixed consumer hardware; (3) being a solo founder limits how fast I can do outreach and engineering at once.
Outcome if it fails: the money still produces a published reference architecture, interview findings and an open demo others can learn from, and I narrow to the one use case operators actually asked for.
$0 in cash. Non-cash support from startup programs: Mixpanel (startup plan), Amplitude, Sentry and Help Scout credits, and an ElevenLabs startup grant (credits) approved on October 7, 2026. Applications pending with Emergent Ventures, EA Funds, Lightcone Commons, BlueDot Impact, the OpenAI Cybersecurity Grant and MIT Solve; I'll update this page if any come through.