No borrowed goals
A continuous health coach for people already testing protocols, wearables, labs, training, and supplements. DN1 asks what outcome you care about, asks whether each signal serves that outcome, then uses the evidence to shape what you do tomorrow.
You're already running an experiment on yourself. DN1 closes the loop.
You notice what no device measures: energy, mood, how you actually feel. Conversation surfaces it as structured data without making you fill out a form.
Your wearable, labs, subjective state, and habits are judged against the outcome you chose. A score can be useful without being in charge.
The evidence record constrains what the coach is allowed to claim. You see the plan and the reasoning, not the math.
Coach surface example from the web app. Not a live production proof claim, and not medical instruction. The useful object is the pattern: one conversation, the relevant signals beside it, and a bounded next move.
“I thought walking and rucking were my cardiac base. 1,145 matched days said the real levers were weight and intensity, so the plan changed.”Mark, founder.
I built DN1 because I needed a system that could tell me when my own protocols weren't doing what I thought they were doing.
Walking moved from “cardiac base” to “anchor practice.” Cardiac training was added separately, on the levers that actually moved cardiac. Rucks dropped Day 41. Norwegian 4×4 intervals added.
Walking stayed in the plan. Not for the heart rate. It stayed because it sets the day's body read, clears the head, and makes harder work more likely. Most tools would have dropped it when the cardiac signal came back flat. DN1 reassigned it to the job it actually performs.
Your version will not be rucks. It might be caffeine, creatine, appetite changes, blood pressure, afternoon fog, or whether a sleep score is actually protecting your day. The job is the same: make the signal prove it serves the goal.
You read your wearable in the morning. You change something. You ask a question. You journal sometimes. A few weeks later you can't remember whether the change mattered. That's the workflow most people are running with their tools. DN1 closes it: capture the intervention, set the verification window, measure how the number changed, keep the read bounded, and remember the context.
Founder dogfood, not market validation. The system has been running on me since March 10. Three proof-of-method artifacts from the first 48 days: what it tracked, what it decided, and the moment it got something wrong and updated.
External pilot ledger: not published yet. Founder dogfood is proof of method, not proof of market.
| metric | rows | days |
|---|---|---|
| energy_level | 49 | 33 |
| mood | 44 | 31 |
| mental_clarity | 40 | 31 |
| motivation | 39 | 28 |
| stress_level | 33 | 24 |
| soreness | 26 | 19 |
| sleep_quality_subjective | 24 | 20 |
| digestion_quality | 21 | 17 |
No forms. No scales. The coach reads the morning braindump and the structured rows fall out the back. Every row keeps the verbatim sentence it came from.
Keep: Thiamax survived stack revisions for energy and clarity.
Drop: Ashwagandha was pulled after morning flatness.
Too noisy: Confounded sleep windows stayed inconclusive.
Updated: Magnesium stayed, but the outcome moved off sleep.
| Intervention | Status | What was tested | Why the call |
|---|---|---|---|
| Magnesium glycinate 420 mg | Reassigned | 14 nights at therapeutic dose, sleep score primary outcome | Sleep null. Vascular and autonomic arms moved. Kept the protocol, moved the verification outcome. |
| Ashwagandha 300 mg bedtime | Dropped | Day 17 morning energy flat against textbook recovery markers | Cortisol blunting carried into the morning. Pulled it. |
| Thiamax (TTFD) 200-300 mg | Kept | Energy and clarity arm, day 13 onward | Strong supportive signal. Survives stack revisions. |
| Nattokinase 12,000 FU | Kept | Part of P-004 BP cluster across 41 days | 146/102 to 121/86, physician-monitored. Stack held. |
| Lactoferrin | Dropped | Lowest impact lever at the start of the window | Cut Day 5. Nothing earned its place. |
| Collagen | Dropped | Skin and connective tissue input | Cut Day 2. Vitamin C supports endogenous synthesis. Redundant. |
| Baobab | Dropped | Mechanism unclear at intake | Cut Day 19. Unclear purpose disqualified it. |
| Extended fasting (24 to 48 h) | Dropped | Liver recovery and autophagy framing | Cut Day 13. Multi-model consensus said net negative for body recomposition. |
| Rucks (45 lb) | Dropped | Apr 7 ruck vs Apr 14 walk, matched pace, paired H10 | Delta 0.8 bpm. Rucks reframed as anchor practice, intervals added. |
| Norwegian 4x4 intervals | Added | Day 41 Cardiac Rebuild Protocol launched | Cardiac stimulus the rucks were not delivering. P-011b active arm. |
| Daily morning oral temp | Added | Cheap continuous monitor for thyroid tone | Day 41 onward. Five dollars a day, replaces a panel as the gating test. |
The drop column is what most products will not show you. Adding to the stack is easy. Pulling something is the harder move, and it is the one that earns the trust.
The system absorbed the correction into how it reasons about me going forward. The product consequence is simple: a generic recovery score can no longer overrule personal evidence in non-clinical training decisions without explaining why. Rest recommendations now require named reason codes; no context-free scalar authority.
The pattern this names: composite recovery products can lean conservative without naming the reason. DN1 should support action only when the user's evidence supports it, and only inside the consumer training and recovery scope.
See all evidence, including the nulls, the noise, and the drops
Three things compound. The founder dogfood data helped sort which levers appeared to move physiology. The conversation captured what mattered beyond the numbers: constraints, tradeoffs, subjective notes, and whether the plan fit the ordinary parts of the day. And the collaboration shaped a plan that survived contact with real life. No single tool we could find closed all three.
What the learning looks like over time, including the things that didn't work.
You show up with something real. “I'm foggy at 3 PM.” “My doctor says my BP is high.” “I want energy left for my kids after work.” The coach asks the next right question. Everything you say is captured, not just the numbers, but what you are trying to make work.
You start building a personal measurement history: morning HR ranges, sleep architecture, and training response. At this stage the system can organize the evidence, not declare a stable baseline.
The system proposes something testable, on your data, not on a population. “Your evening HRV is consistently low. Let's test whether cutting caffeine after 2 PM changes that over the next ten days.”
Some interventions may have enough measurements to keep, drop, or reassign. At least one may still be too noisy to call. The system says that out loud and names what would close the question.
You tell the system you'd rather do push-ups between calls than block 45 minutes for the gym. The plan can be revised around that constraint. You miss a day. The system keeps the context, protects the anchor practice, and drafts the next executable move without pretending the original plan still fits.
Most people in this audience already spend $200 to $500 a month on supplements, devices, peptides, lab panels, sometimes a coach. The cost isn't the question. The question is which of those parts is earning its place.
Paid pricing and refund terms are not live yet. The current public position is a planned offer, not an active price or guarantee claim.
We're inviting users in cohorts. Tell us what you're trying to figure out about your body and we'll be in touch when it's your turn.