Three free general-purpose chat tools, used in order, become one coordinated personal-health intelligence system. No new app. No subscriptions stacked on subscriptions. A method that survives any platform, any model, any vendor.
The three layers
Each layer has one job, and you pick the tool that does that job well today. Together they form a closed loop: evidence in, lived data through, protocol out, lived data back in. Method over brand — when a tool changes, the architecture still works.
Layer 01 · Sourced-search AI
Read the actual literature on a single question — RCTs, meta-analyses, mechanism — with citations you can click.
Output
A living, sourced reference library tuned to your body — not another bookmark folder.
In practice
Magnesium + statins → 2-hour gap. Ginger + blood thinners → caution. B12 + PPIs → low-absorption risk. Each claim cites its source.
Layer 02 · Long-context AI
60-second daily logs accumulate into months of context — sleep, cycle, food, mood, training, labs — held in one continuous conversation.
Output
A chronological, synthesised biological narrative — your story, in your words, kept by you.
In practice
Three logs a week is sustainable. Insight density peaks around week three. Patterns the app can't see start to surface.
Layer 03 · Conversational AI
Translate Research evidence + Ledger patterns into one calm, single-variable experiment that fits the actual week ahead.
Output
A living protocol you can defend to a practitioner — not a textbook fantasy you'll abandon by Wednesday.
In practice
Most generic protocols fail the feasibility test. Constraint-aligned protocols compound — quietly, weekly.
The 7-Intelligence Flow
The pipeline a question travels — from the kitchen, through the stack, back to your plate.
A real moment. A symptom, a question, a curiosity.
Research layer pulls evidence with mechanism + caveat.
Evidence hierarchy applied. Hype quietly demoted.
Ledger entry — 60 seconds, plain language.
Recurring signals named after 5 days of data.
Sunday Integration Hub: pattern + evidence merge.
A protocol that fits your week, not someone else's.
Evidence Hierarchy
Every research output is ranked before it earns a place in your protocol. RCTs and meta-analyses outrank influencer claims — transparently, every time. We label the tier; we never hide it.
Strong
RCTs · Meta-analyses · Cochrane reviews
Promising
Mechanistic studies · Rodent data · Small trials
Anecdotal
Influencer claims · Single-n stories · Marketing
Pattern Signals
We don’t name a pattern from one log. By Day 5, recurring signals start to repeat — and only then do we mark them.
5 days
of data
The Sunday Integration Hub
Daily perfection breaks. A weekly hub holds. Sunday is the anchor that keeps Research, Ledger, and Protocol speaking to each other.
What did this week's evidence change? Which beliefs got upgraded, which got retired?
Which signals repeated? Which were one-offs? Which deserve a hypothesis next week?
What in the current plan still fits real life? What should be smaller, slower, or stopped?
Prompt Anatomy
When you see the formula, you stop writing prompts and start engineering them.
Input
Age, visceral fat, allergies, current meds (e.g. GLP-1), known conditions.
Input
HRV, sleep score, last 24h food log, recent symptoms or wins.
Input
Harm reduction, food sequencing, somatic reset, schedule, budget.
Output
= An actionable, life-fitting recommendation. Reviewable by a practitioner.
For practitioners
The Practitioner Fork uses the same 3-layer architecture and produces a Session-1 Intelligence Package — so the first session starts where the fifth usually does, without losing care quality.
Replace 3–5 sessions of baseline gathering with a coherent, client-owned narrative the practitioner can review in minutes.
The stack carries cognitive load between sessions. You stay the practitioner; the architecture handles the continuity.
The free 10-day challenge runs the full method on your real life — one small step at a time. You keep what you build.
Still exploring? See what's suggested for you →