Weights see. Substrate knows. Baselines personalize. Policy decides how much to say.
The platform reads the body through several specialized modules — “hands” — for movement, tongue, nutrition, hydration, metabolism, body maps and, next, vision and jaw. A central brain connects what the fingers find. The temptation in 2026 is to hand all of this to a large AI model. This page is an overview of where our research has led, and it happens to argue the opposite.
Almost everything that makes the platform valuable — practitioner knowledge, compensation maps, the relationships between body systems, the rules for how cautiously to speak — is substrate: written, readable, citable knowledge. Learned model parameters — weights — are needed in only one place: turning camera pixels into measurements. Between the two sits a layer that makes the platform personal: each person’s own baseline, fitted from their captures over time. Language models are kept at the edge, as assistants that may suggest but never decide.
“AI” is used loosely for three very different tools
Each has a job it is good at and a cost it carries.
Automation
Rules, formulas and lookups that someone wrote down.
Use when the right answer can be written down in advance.
COST · AUDITInstant and free. Every step readable.
Machine learning
Parameters learned from many labelled examples.
Use when the input is raw signal — pixels, video — or there are thousands of labelled outcomes.
COST · AUDITFree on device once trained. Testable, not readable.
Language AI
A pre-trained model that reads and writes text.
Use when the input is unstructured text and something else checks the output.
COST · AUDITSeconds to minutes, or paid per call. Neither readable nor reliable on its own.
Our rule of thumb. If you can write the test with an exact expected answer, it should be a script. A language model never judges another model — the referee is always a script.
Ontology
A shared vocabulary for what the system reasons about
The more research a person does in the medical field, the more they see an intentional separation of vocabulary for things that can’t be argued apart: the same, or at least intimately connected and synergistic.
Every term in the diagram exists in code today.
The epistemic layer
How the system knows what it knows, and how sure it may sound
To account for bias we work at the level of epistemics. Health and wellness is a field where well-evidenced practitioner observation is often absent from mainstream sources; the platform must neither suppress it nor overstate it.
Responsibilities & research protocol
What we follow, and require of anyone who uses our services
These are the questions we take to regulatory counsel and an ethics board. The answers below are ours, not theirs.
Claims
Staying a general-wellness product depends on never claiming to diagnose, treat or prevent a condition — in the app, in marketing and in model output.
Research ethics
Collecting people’s data to establish or validate correlations is human subjects research: informed consent, and ethics review to publish or partner with clinics.
Health data
HIPAA if you are or serve a covered entity; consumer health-data and breach rules regardless. Encryption, access control and retention are yours to get right.
Measurement validity
Show measurements match ground truth, repeat reliably, and that practitioners agree on what they label.
Knowledge provenance
Keep practitioner observation labelled as such — distinct from clinical evidence — everywhere it surfaces.
Research
The headlines. Full papers on request.
Each piece below is summarized at the surface. Published work links straight to its DOI. Drafts that aren’t published yet are shared in full, with methods and limits, on request.
Food · Nutrition policy
White paper draft v1.01 · Not peer reviewed
Dense energy isn’t usable energy
From the manuscript Destroying Food Myths and the white paper Reevaluating Nutritional Policy Through Molecular, Ecological, and Institutional Lenses.
4 of 27vitamins and minerals with a Daily Value that a US label must list
−84%magnesium in white flour vs whole grain; enrichment doesn’t put it back
63%of graded catalog products are ultra-processed (NOVA 4)
Calories measure how much energy a food holds, not how much the body can use. Usable energy runs on micronutrients.
Public frameworks — calories, PDCAAS, the food pyramid — were built under constraints that no longer hold, and neglect gene–nutrient interactions, trace minerals and the microbiome.
People aren’t all the same, and neither are soil, water or access. Guidance should measure, not assume.
The findings that get discussed as a diet story apply to what goes on skin and hair — and that part is rarely said out loud.
Industry shaping what the public hears is a familiar story in nutrition. The same pattern has played out in personal care, told as one-off product scandals rather than a pattern.
Absorption is discussed in terms of digestion. The same measurements exist for skin.
Who funds ingredient safety data, and how long findings take to surface, matters as much on the shelf as on the plate.
Movement · Fascia · Assessment
Published · Zenodo · v0.1.0 · Open access
Superior Terminus Framework
Movement assessment, fascial chain correspondence, and the case for multi-terminal triangulation body assessments. An independent evidence synthesis by Joshua Harris, published 17 September 2026 (CC BY-NC 4.0).
Three terminals — the feet, the spinal column and the tongue — read at the same moment and in the same position produce a pattern none of them surfaces alone. “The claim is triangulation, not direct correspondence.”
The tongue is the superior terminus of the deep front line and one of the five diaphragms of the osteopathic model. Oral assessment is non-invasive and readable anywhere.
Non-athletes spend most of the day seated or reclined, yet assessment treats upright posture as the functional baseline. Position is the missing modifier.
Status: technically feasible, not yet executed. Assessment protocols are at the pattern-surfacing stage; inter-rater reliability has not been established.
Our research is anchored in work anyone can review. Each card links straight to the paper by its DOI, with a quote taken word for word from the published abstract.
Movement, fascia & the tongue
“A total of 92 (31.1 %) students were aware of OMT as a modality in treating musculoskeletal disorders.”
Awareness and interest in osteopathic manipulative treatment in allopathic medical studentsDarby A, Parascando JA, Lipinski M, et al. · Journal of Osteopathic Medicine 2023;123(8):379–384 · Open access (CC BY 4.0)Why it matters: knowledge one tradition of medicine uses is unknown to most students of the other — the vocabulary gap our ontology is built to bridge.doi:10.1515/jom-2022-0232 ↗
“Currently, there is no dynamic non-instrumental test in the scientific literature to highlight any lingual dysfunctions.”
Non-Instrumental Test for the Evaluation of Tongue FunctionBordoni B, Escher AR · Cureus 2021;13(9):e18333 · Open accessWhy it matters: the gap the Superior Terminus Framework’s tongue assessment sets out to fill.doi:10.7759/cureus.18333 ↗
“The five diaphragms (tentorium cerebelli, tongue, thoracic outlet, thoracic diaphragm and pelvic floor) represent an important tool for the osteopath to evaluate and find a treatment strategy…”
The Five Diaphragms in Osteopathic Manipulative Medicine: Myofascial Relationships, Part 1Bordoni B · Cureus 2020;12(4):e7794 · Open accessWhy it matters: places the tongue in anatomical continuity with the pelvic floor.doi:10.7759/cureus.7794 ↗
“This study validates the TRMR-LPS as a useful functional metric for assessment of posterior tongue mobility.”
Assessment of posterior tongue mobility using lingual-palatal suction: Progress towards a functional definition of ankyloglossiaZaghi S, Shamtoob S, Peterson C, et al. · Journal of Oral Rehabilitation 2021;48(6):692–700 · 611 subjects, ages 3–83Why it matters: a measured, repeatable way to read tongue mobility.doi:10.1111/joor.13144 ↗
“A weak but significant correlation was observed between TMD and the degree of ankyloglossia, particularly when assessed using Kotlow’s method.”
Anatomical versus functional classification of ankyloglossia and their association with temporomandibular joint disorders in adultsBeckmann A, Peroz I, Peroz S · Journal of Oral & Facial Pain and Headache 2025;39(3):77–83Why it matters: tongue restriction and jaw dysfunction travel together in adults — and the authors call for a validated adult assessment tool.doi:10.22514/jofph.2025.049 ↗
Food & nutrition
“The Atwater factors, when applied to almonds, resulted in a 32% overestimation of their measured energy content.”
Discrepancy between the Atwater factor predicted and empirically measured energy values of almonds in human dietsNovotny JA, Gebauer SK, Baer DJ · American Journal of Clinical Nutrition 2012;96(2):296–301Why it matters: the calorie on the label describes the food, not the body that digests it.doi:10.3945/ajcn.112.035782 ↗
“The SRF’s funding and role was not disclosed.”
Sugar Industry and Coronary Heart Disease Research: A Historical Analysis of Internal Industry DocumentsKearns CE, Schmidt LA, Glantz SA · JAMA Internal Medicine 2016;176(11):1680–1685Why it matters: who funds the evidence shapes what the public is taught — in food, and on the personal-care shelf.doi:10.1001/jamainternmed.2016.5394 ↗
All peer-reviewed sources cited by the Superior Terminus Framework
Bayattork M, et al. Exercise interventions to improve postural malalignments in head, neck, and trunk. J Exerc Rehabil 2020;16(1):36–48. doi:10.12965/jer.2040034.017
Salsali M, et al. Association between physical activity and body posture: a systematic review and meta-analysis. BMC Public Health 2023;23:1670. doi:10.1186/s12889-023-16617-4
Paillard T. The optimal method for improving postural balance in healthy young and older people. Front Physiol 2023;14:1188496. doi:10.3389/fphys.2023.1188496
Chen Y-L, Zhang L-P. Postural variability in sitting: comparing comfortable, habitual, and correct strategies across chairs. Appl Sci 2025;15(13):7239. doi:10.3390/app15137239
Cohen RG, et al. Lighten Up! Postural instructions affect static and dynamic balance in healthy older adults. Innov Aging 2020;4(2):igz056. doi:10.1093/geroni/igz056
Bordoni B, Escher AR. Non-instrumental test for the evaluation of tongue function. Cureus 2021;13(9):e18333. doi:10.7759/cureus.18333
Deregibus A, et al. Analysis of tongue function from the orthodontist’s point of view. Appl Sci 2021;11(6):2520. doi:10.3390/app11062520
de Andrada e Silva M, et al. Postura, tônus e mobilidade de lábios e língua de crianças respiradoras orais. 2012. doi:10.1590/S1516-18462012005000002
Maeo K, et al. Low tongue posture improvement effect of orofacial myofunctional therapy. Orthod Craniofac Res 2025;28(5):872–881. doi:10.1111/ocr.70000
Zaghi S, et al. Assessment of posterior tongue mobility using lingual-palatal suction. J Oral Rehabil 2021;48(6):692–700. doi:10.1111/joor.13144
Tang Y, et al. Ultrasound quantification of tongue-palate distance and tongue thickness in TMD patients. BMC Oral Health 2025;26(1):156. doi:10.1186/s12903-025-07568-w
Bordoni B. The five diaphragms in osteopathic manipulative medicine: neurological relationships, part 1. Cureus 2020;12(6):e8697. doi:10.7759/cureus.8697
Bordoni B. The five diaphragms in osteopathic manipulative medicine: neurological relationships, part 2. Cureus 2020;12(6):e8713. doi:10.7759/cureus.8713
Bordoni B. The five diaphragms in osteopathic manipulative medicine: myofascial relationships, part 1. Cureus 2020;12(4):e7794. doi:10.7759/cureus.7794
Beckmann A, Peroz I, Peroz S. Anatomical versus functional classification of ankyloglossia and TMJ disorders in adults. J Oral Facial Pain Headache 2025;39(3):77–83. doi:10.22514/jofph.2025.049
Deshkar M, et al. The influence of the tongue on the development of dental malocclusion. Cureus 2024;16(5):e61281. doi:10.7759/cureus.61281
Sigal A. Changes in tongue resting posture following pediatric lingual frenotomy. Int J Orofac Myol Myofunct Ther 2026;52(1):9. doi:10.3390/ijom52010009
Also cited without a DOI: Braune & Fischer, Das Gehen des Menschen (1895); Elftman, J Exp Biol 1934; Myers, Anatomy Trains.
Request a paper
Read the full document
Tell us who you are and which paper you’d like. We share full drafts with researchers, practitioners, reviewers and partners.
Thanks — we’ve received your request and will reply by email.
Currant · The food research in an app
Everything a label leaves out, in one scan
Scan any product
Point your camera at a barcode, or search millions of foods by name.
Nutrition Score, 0 to 100
How whole, fibrous, hydrating and clean a food is. The same score for everyone, whatever your goal.
Quick vs lasting energy
Whether a food hits fast or carries you for hours, scored separately so goals never bend the quality score.
Harmful additives flagged
A carcinogen or harmful additive caps the score, instead of being averaged away.
Better alternatives
When a product scores low, Currant shows similar products that score higher.
Recalls and allergen watch
Food recalls, and the allergens you choose to watch for.
Most food apps count. Currant measures.
MOST FOOD APPSCount calories and macros.CURRANTLooks at minerals, vitamins, oils and enzymes: the part that decides whether energy is usable.
MOST FOOD APPSShow what’s on the label: 4 micronutrients.CURRANTFills in the rest from USDA composition data, per 100 g so foods compare fairly.
MOST FOOD APPSGrade protein by grams per serving.CURRANTGrades protein quality with DIAAS, the method FAO recommended in 2013.
MOST FOOD APPSPush a diet or a meal plan.CURRANTWorks with how you already eat, whether traditional, USDA, CDC or your clinician’s plan.
MOST FOOD APPSSell placements and ads.CURRANTIndependent: no brand or manufacturer can pay for a rating, and there are no ads.
One free account for Currant and the LHP apps. Optional — Currant works fully without one.
