03 · Personal Brain Install
I build you an AI that knows your work — and belongs to you.
Fixed-scope project · 2–4 weeks from kickoff to delivery. Pre-Conway founding rate: $4,500. 30-day post-install support included.
Solo professionals, knowledge workers, operators of single businesses. The person who knows the role and is going to own the brain.
Hand me your files. I turn them into a private, searchable brain you can call up from Claude, ChatGPT, Cursor, or anything modern. Gated to you. Portable when you leave.
The argument for owning your own brain.
Every quarter, the AI vendors get smarter about you. ChatGPT now remembers across chats. Claude has projects that retain context across conversations. Gemini reads your Workspace. The trend line is obvious: within a year, one of those vendors will ship a continuously-aware assistant that knows you the way a coworker who’s been with you for five years knows you.
The question isn’t whether that’s coming. It is. The question is whether, when it lands, your relationship with your work is structured to take advantage of it — or whether you have to start the relationship over with whichever vendor wins that race, on whatever terms they set.
The Personal Brain Install is the answer to the second version of that question. It’s a real, working AI brain — built from your actual files, structured around the way your role actually thinks, sitting on a subdomain you brand, gated to you alone — that any modern AI tool can query and benefit from. It’s not a chatbot. It’s the corpus underneath the chatbot, in a format you own.
When the next model launches, your brain plugs into it the day it ships. When you change jobs, your brain comes with you. When the company you work for adopts a different AI vendor, your brain doesn’t care — it works with all of them. The thesis from the homepageis "I’m the human between you and every AI vendor." The Personal Brain is the artifact that proves it.
The diagnosis
Three things break without one.
Your work is locked inside someone else’s product.Every Claude project, every custom GPT, every Notion AI workspace is investment into a system you don’t control.
Each new AI tool starts from zero on you. You re-teach every tool what the last one already knew, and most people give up before the tool proves itself.
The retrieval shape is wrong by default. Tools give you a vector database when you need a document tree. The shape matters more than people think.
“When the AI giants ship continuous memory, the people who already have their work organized in their own brain are the ones the new tools learn from — not the other way around.”
What the install builds
Files in. Brain out. Yours.
You hand me your files — documents, writing, projects, saved chat exports, case studies, deliverables. I extract the same four layers from the Audit and structure them into a corpus organized for retrieval the way your work flows.
I deploy that corpus to a private, branded subdomain gated by Cloudflare Access. It exposes an MCP endpoint — the standard protocol every modern AI tool uses to query external knowledge. Claude, ChatGPT, Cursor — any tool that supports MCP can query your brain.
Three to five role-tuned skills get installed inside your daily AI tool. A capture surface (Telegram, Shortcut, or web form) keeps the brain current. The retrieval contract is shape-matched to your role — vector where vector belongs, document-tree where it belongs, never one-size-fits-all.
The window is now
Build the brain before the vendors do.
$4,997 fixed. Two to four weeks. Yours when you change jobs, change tools, or the next model launches. The Diagnostic Call scopes the install in 30 minutes.
How a four-week install runs
The cadence.
- W1
Diagnostic + corpus pull
Kickoff call. File access. Four-layer extraction. Draft corpus by Friday.
- W2
Architecture + retrieval
Retrieval shape designed to your work. Subdomain, Cloudflare Access, MCP server, capture surface deployed.
- W3
Skills + integration
3–5 role-tuned skills installed in your daily AI tool. You start using the brain in low-stakes work.
- W4
Hardening + delivery
Edge cases, docs, handover call. Source files, schema, deployment — yours outright. 30-day support window opens.
What you walk away with
In the bundle.
A working brain
Live, gated, branded, queryable from any AI tool you already use. The brain shows up inside Claude or ChatGPT as another piece of context.
Source files you own
Corpus in Markdown, schema in plain text, deployment config in a private repo. Rehost it yourself in a weekend if you ever leave.
Skill library
3–5 reusable workflows tuned to how you actually work. Named, documented, modifiable. Yours.
Portfolio surface
The brain doubles as a portfolio piece — take it to a pitch, a next role, a board meeting. Structured to be showable.
30-day support
Bugs, tweaks, schema adjustments, additional skill installs — included for the first month, no extra invoice.
Retainer option
Optional ongoing care: Light $497/mo or Standard $997/mo. Not required — most installs are stable on their own.
The math
Sub-year payback
~20 hours of skilled work at $150/hr effective + productized infrastructure. For any knowledge worker saving 5–10 hrs/week, the Install pays itself back inside a year — before the Conway-style continuously-aware assistants land.
Volume bands — per-employee pricing
When you install for a team.
Solo · 1 brain
$4,997. The single-operator install.
Small · 2–9
$4,997 each. Architectural work shared across the cohort.
Mid · 10–19
$4,497 each. Cross-brain query rights become economical.
Growth · 20–49
$3,997 each. Institutional patterns codified.
Org · 50–99
$3,497 each. Department-level governance in scope.
Enterprise · 100+
Custom. The institutional architecture is the deliverable.
Thinking team-scale?
The Company Install is the next page over.
Per-employee Personal Brains paired with a shared Company Brain that holds institutional knowledge. Read the architecture, then we talk.
Who this is for
Named by role.
- The senior knowledge workerConsultant, analyst, principal IC. Your work product is your thinking, captured across hundreds of documents. You want that thinking compounding instead of evaporating.
- The founder or solo operatorYou run a business of one or a small business with you as the brain. Your corpus is your edge. You want it portable and queryable.
- The career-stage professionalYou’re between roles or planning to be. You want to take your AI training with you and have something concrete to show the next room.
- The agency principalYou've been refining methodologies, frameworks, and client patterns for years. They live across Notion, Drive, and a Claude project. The Install consolidates them into a single brain you own outright.
What’s included
In the price.
- Diagnostic Call → file audit → four-layer extraction → corpus build
- Shape-matched retrieval contract (vector, document-tree, tabular, graph as your role requires)
- Branded subdomain (yours.brandedmayhem.com or your domain), Cloudflare Access gated to you
- MCP server — Claude, ChatGPT, Cursor, or any modern AI tool can query the brain
- 3–5 role-tuned skills installed inside the AI tool you already use
- Capture surface — Telegram, iOS Shortcut, or web form for ongoing input
- Demonstrated-capability portfolio surface — the brain doubles as a showable artifact
- Claude for Small Business setup when applicable
- 30-day post-install support — bugs, tweaks, schema adjustments included
- Full ownership of source files, schema, and deployment config
Common questions
Asked, answered.
How is this different from a Claude project or custom GPT?
A Claude project or custom GPT lives inside one vendor’s product. The Personal Brain lives at a URL you own and answers to every vendor’s tool. If Claude raises prices or changes a feature, your project is affected. If your brain’s hosting provider goes down, you move it to a different host without losing data. The artifact is different in kind.
The other difference is the corpus structure. A Claude project holds files but does basic retrieval on them. The Install applies a shape-matched retrieval contract — vector where vector belongs, document-tree where document-tree belongs, tabular where tabular belongs. That’s the work that almost nobody does by default.
What happens to the brain if I stop working with you?
You keep it. You own the source files, the schema, the deployment config, and the subdomain. You can rehost it yourself, hand it to another developer, or let it run as-is for as long as you want. There’s no lock-in. The retainer is optional and exists for people who want me to keep tuning — not as a requirement to keep using what you bought.
What are the pass-through costs?
Real but small. Cloudflare Workers + D1 + Vectorize + Workers AI run typically $0–$20/month per brain depending on usage — included in the Cloudflare Free tier for most personal use, single-digit dollars even for heavy use. The custom subdomain is $1–$2/month amortized. If you choose Telegram or SMS capture, Twilio adds $5–$20/month. LLM tokens, when applicable, are at-cost from your Anthropic or OpenAI account.
Most personal brains run for $5–$30/month all-in. The infrastructure cost isn’t the lock-in mechanism.
Why $4,997?
Roughly 20 hours of skilled work at $150/hr effective, marked up for the productized infrastructure (Workers, D1, Vectorize, Workers AI, Access deploy, branded subdomain, MCP server, capture surface), 30-day support, and the productized risk premium. Sub-year payback for any knowledge worker saving 5–10 hours/week. The math holds even before the new-AI-tool advantages compound.
What's the pre-Conway founding rate?
$4,500 — $497 off — locked in for the first 25 installs OR until one of the major AI vendors (Anthropic, OpenAI, Google) ships a continuously-aware coworker assistant, whichever comes first. The naming is a wink at Star Trek’s persistent-Conway memory — the moment the AI vendors cross that threshold, the install becomes more valuable, not less, because the new tools learn from your existing brain rather than competing with it. Founding-rate buyers keep $4,500 on retainer renewals.
Can I bring the brain to my company?
Yes, and that’s the design. The brain is portable across employer changes because it’s yours, not your employer’s. If your next company adopts AI tooling that supports MCP (most modern stacks do), your brain plugs in on day one. If they don’t, the brain still works for you privately. The career-portability is a feature, not an accident.
How technical do I need to be?
You don’t need to be technical. The whole point of the productization is that the technical work happens on my side. You provide files, you give feedback on the shape, you use the brain through tools you already know. If you’re technical and want to dig into the deployment config or the retrieval architecture, the source is yours and I’ll walk you through it — but that’s optional.
What if my files are confidential or regulated?
I sign NDAs. The infrastructure runs on Cloudflare with Access gating — everything in transit is encrypted, everything at rest sits on infrastructure you control through your own account if you prefer. For regulated work (legal, medical, financial), we’ll talk through the specific compliance requirements during the Diagnostic Call and adjust the architecture to match. I’ve built installs for engagements with HIPAA-adjacent content; the infrastructure scales to most reasonable requirements.
Adjacent paths
Where this connects.
Start with the Diagnostic Call.
The Install starts with a 30-minute Diagnostic Call. Free. Scoped to figuring out whether the Personal Brain is the right move for you, or whether the Audit is — or whether neither is, and we should be talking about something else entirely.
As featured in
