Hire me to automate the work that drains your week, repair the app that is already live, build the site or product itself, or train an operator agent that works inside boundaries you control.
Every engagement starts through the same $150 diagnosis. I inspect the real system, write down what I found, and price the next step before work begins.
“I am paying people, or paying myself in hours, to do things a computer should already be doing.”
“People are using it, something is broken, and I need the real answer — not a rebuild pitch.”
“I want to talk to the person building it, own the result, and know the price before work starts.”
“I cannot do this every day, but I am not handing my business to a robot I cannot watch.”
This is a working diagnosis, not a sales call. You leave with a written plan, a real scope, and a fixed price whether or not you hire me. If we build together, the full fee is credited to the work.
The work begins with your actual business, not a generic discovery worksheet.
You show me the system, workflow, or idea. I trace the problem and write down what is worth doing first.
You receive the deliverables, timeline, ownership terms, and a fixed price. No step starts before you approve it.
I drive the setup. You only log in when I point at a screen. You review a working preview before anything ships.
I am Frank Epps. I design and build the work myself: the interface, the integrations, the safety boundaries, the deployment, and the documentation you keep afterward.
Across 26 working systems, I have built AI operations, products, knowledge tools, real-time data interfaces, business automation, and interactive worlds.
The complete capability menu, personal-build screenshots, and counted inventory now live on one focused page.
Review capabilities & workI run a governed autonomous agent system on my own business. The Orrery organizes work across three leadership lanes, the Armillary exposes evidence and approval points, and a read-only overseer performs weekly cross-domain synthesis. The Personal App Engine proves the same default-deny, sandboxed approach for building software. Human approval remains the release boundary.
The platform, command console, measured evidence, and application-building boundaries now have their own page.
See the systems →Three examples of diagnosis, delivery, and production work.
The reported failure looked like matching logic. The real cause lived in empty data. I diagnosed it, rebuilt the engine, verified it, and shipped it to production.
The build connected the customer-facing site to workflows for posting, booking, review requests, and rebooking — designed so the owner only logs in when pointed at a screen.
A complete subscription SaaS with authentication, tiered payments, row-level security, serverless APIs, and AI-generated output.
No hourly meter. No silent scope expansion. Larger products and operator agents are priced after the diagnosis because their scope genuinely varies.
| Stage | Price | What it means |
|---|---|---|
| Diagnosis | $150 | A written automation map, technical finding, or scoped build plan. Fully credited if we continue. |
| Live app or site rescue | $800–1,500 typical fix | Fixed after the diagnosis. Rebuilds or migrations are quoted separately. |
| Small-business site + AI | $1,500–2,500 | A production site and its scoped AI partner. Larger products are re-quoted in phases. |
| Operator agent build | Scoped after diagnosis | A recurring business function with approval gates, escalation, auditing, and supervised handoff. |
Tuning, safety monitoring, and reporting are available and scoped after the build. You keep the agent, code, accounts, and operating knowledge either way.
The diagnosis exists to reduce uncertainty, including the possibility that the right answer is not to build.
I say so. You still keep the written finding and the practical next step. The $150 buys the answer, not a commitment to hire me.
Yes. The code, accounts, domain, and operating documentation are yours. You receive a walkthrough at handoff.
I use explicit permissions, approval gates, spending caps, a kill switch, an append-only action log, and a separate checker. I do not ship unattended money-spending autonomy.
No. Ongoing tuning, monitoring, and reporting are available and scoped after the build. You keep the agent, code, and accounts either way.
The onboarding promise is one 20-minute call. I drive; you log in only when I point at a screen and review the working result before release.
Often, yes. I have worked across custom code and no-code systems. The diagnosis determines whether a direct repair, integration, or migration is the honest path.
Book the $150 diagnosis now, or leave a short note if you are not ready to pay today. I will read it myself.
Book the diagnosisUse the guided tool to organize what you need. Any estimate is a starting point; the paid diagnosis produces the fixed scope.
A starting estimate, not a final scope.
