Medical Sources & Limits

Public source registry for Ready Trainer One — the primary references behind RTO's health information, with status labels and explicit limits.

How to read this page

Each entry below is tagged with a status label so you can tell at a glance how that source is used in the product:

Draft-only papers and pre-prints are not listed as accepted sources. If a draft becomes a final accepted reference, it moves into the relevant section with an explicit status label.

AI health literacy and patient agency

How patients, clinicians, and tools should think about AI in health contexts. This framing shapes how RTO presents AI output — small, hedged, source-cited, and never positioned as a diagnosis.

• Core source National Academy of Medicine — Critical AI Health Literacy as Liberation Technology Perspective (Campos et al.). Anchor for RTO's posture: surface AI outputs as one input the patient interrogates, not as an authoritative reading. Campos, H., and L. Salmi. 2025. Critical AI health literacy as liberation technology: A new skill for patient empowerment. NAM Perspectives. Commentary, National Academy of Medicine, Washington, DC. https://doi.org/10.31478/202512a

Training and activity safety

What "active enough" and "safe to push harder" mean at a population level. RTO never prescribes exercise; these references frame the conservative defaults the coach reasons from.

• Core source CDC — Physical activity basics: adults U.S. patient-facing physical-activity guidance: aerobic and muscle-strengthening minimums for adults. The plain-language version we cite in chat when explaining a baseline.
• Clinical reference Quantity and Quality of Exercise — ACSM position stand American College of Sports Medicine. Medicine & Science in Sports & Exercise, 2011. Frequency, intensity, time, and type prescriptions for developing and maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness in apparently healthy adults — the professional-society anchor behind RTO's training-load defaults.

Sports nutrition and fueling

How an active person — including someone on a GLP-1 with a reduced appetite — should think about fueling for training. RTO is intentionally cautious here: the interpretation gap between general-population nutrition guidance and athletic fueling is one of the places where personalized AI output can go most wrong.

• Clinical reference Nutrition and Athletic Performance — joint position stand Academy of Nutrition and Dietetics, Dietitians of Canada, and American College of Sports Medicine. Medicine & Science in Sports & Exercise, 2016. Energy availability, protein timing, and carbohydrate ranges for trained athletes — the professional-society anchor behind RTO's fueling framing.

Sleep and recovery

How much sleep adults need and what "recovery" means at the population level. RTO surfaces sleep duration and continuity as inputs to readiness; it does not score sleep quality clinically.

• Core source CDC — About sleep Includes the CDC's adoption of the joint AASM / Sleep Research Society adult sleep-duration recommendation (7+ hours per night). Anchor for the readiness model's sleep-floor logic.
• Core source MedlinePlus — Healthy sleep NIH / National Library of Medicine patient-facing overview. Used for plain-language explanations in chat.

GLP-1 medication context

Drug labels and patient-facing references for the GLP-1 receptor agonists most commonly seen in RTO users. Used as the source-of-truth for storage, common side effects, dose-timing context, and what RTO will never touch (the prescription itself).

• Clinical reference Semaglutide (Ozempic, Wegovy) — MedlinePlus drug information NIH / National Library of Medicine, drawn from the FDA-approved label. Used for warnings, common side effects, and storage guidance referenced in chat.
• Clinical reference Tirzepatide (Mounjaro, Zepbound) — MedlinePlus drug information Same source posture as semaglutide. RTO never changes your prescription, dose, or schedule.

Hydration and electrolytes

Hydration matters more on a GLP-1 (reduced thirst cues, GI side effects). RTO can prompt and log fluids, but it does not replace urgent-care judgment.

• Core source Dehydration — MedlinePlus NIH / NLM. Signs, severity bands, and when to seek care — the language RTO defers to when nudging about fluids.

AI data handling and platform privacy

The terms governing what the AI provider may do with your requests, and the platform privacy posture RTO depends on. Both inform the consent copy in the app and §3 of the Privacy Policy.

• Core source Anthropic — Commercial Terms of Service The Customer-Content / no-training clause RTO relies on for managed and BYO requests.
• Core source Anthropic — Data Processing Addendum DPA terms for customer content processed through the API.
• Core source Anthropic — How we protect personal data (plain-language summary) Plain-language complement to the legal terms above.
• Core source Apple Developer — Protecting user privacy (HealthKit) Apple's on-device privacy posture for HealthKit data. RTO inherits this; raw HealthKit never leaves the phone.
• Core source Apple — Human Interface Guidelines: HealthKit Permission-prompt and data-presentation expectations RTO follows.

Candidates under review

Sources we're tracking but have not yet adopted into a guardrail. Listed here for transparency; do not infer endorsement.

• Candidate under review An Overview of Heart Rate Variability Metrics and Norms Shaffer & Ginsberg, Frontiers in Public Health, 2017. RTO surfaces HRV as one signal alongside sleep and resting heart rate. Interpretation norms in this field are contested; we have not yet pinned a clinical reference for HRV-based readiness logic, and we will not until one we're confident in becomes available.

What these sources do not mean

• The presence of a citation here does not mean RTO's personalized output has been clinically validated. The sources frame the field; the AI's reading of your particular data has not been independently studied.
• None of these sources sanction RTO as a medical device, a diagnostic tool, or a treatment.
• None of these sources authorize RTO — or you, on RTO's advice — to change a prescription, dose, or treatment plan. That conversation is between you and your prescriber.
• A "Core source" label means the source informs a product guardrail. It does not mean the source has reviewed RTO.
• A "Candidate under review" label means we are not yet relying on the source. It is listed to be transparent about what we are reading, not to claim it as evidence.
• AI output can be wrong or incomplete on any given request. Always check with a clinician before acting on a health-related answer.
• AI-generated content — Coach answers, Daily Brief, Today Hero, and food-parse outputs — can be wrong or incomplete on any given request, even when the underlying framing is drawn from a Core source. Treat the model's specific reading of your data as a starting point, not an authoritative reading.