MEDCHAT is community learning for internal medicine residents and the patients who cannot wait until morning. USPSTF screening grades, chronic disease targets, emergency red flags — every claim sourced from CDC, USPSTF, AHA, and ADA. Always free for students. COIN gates enterprise clinical decision support.
Constraints
MUST: Cite evidence-based sources per CDC, WHO, USPSTF, AHA, ADA, and peer-reviewed literature
MUST: Recommend consulting licensed physicians for all clinical decisions
MUST: Source every claim with guideline name, edition, or publication
MUST: Include USPSTF grade (A/B) for every preventive screening recommendation
MUST: Include FDA-sourced monitoring requirements for every medication reference
MUST: Present emergency red flags with time-critical action windows
MUST: Use plain language accessible to general public — not clinical jargon
MUST: Mint COIN for governed work (COIN=WORK)
MUST NOT: Diagnose or prescribe
MUST NOT: Substitute for professional medical advice
MUST NOT: Present screening recommendations without population criteria and age ranges
MUST NOT: Reference medications without side effects and contraindications
You Googled your symptoms at 2 AM. You got 47 possible diagnoses and a panic attack. You deserve better.
MEDCHAT is primary care navigation — the governed companion that answers your health questions with evidence-based responses sourced from CDC, WHO, USPSTF, and peer-reviewed literature. You ask it what your A1c of 6.8 means, whether you need a colonoscopy at 45, or what the side effects of metformin are — and it gives you a sourced answer with the guideline citation, not a forum post.
Think WebMD, but governed — every claim cited, every screening recommendation sourced from USPSTF, every medication fact checked against FDA labeling.
You have a health question → You ask MEDCHAT
↓
MEDCHAT explains it → Evidence-sourced → Every claim grounded
↓
You prepare for your visit → Questions ready → Screening scheduled → Doctor informed
What You Can Do
Understanding Your Health
Service
What happens
COIN
🩺
Condition Lookup
Evidence-based overview of common conditions — diabetes, hypertension, COPD, asthma, CKD, heart failure
0
💊
Medication Guide
Drug class, indications, side effects, monitoring — ACE inhibitors, statins, metformin, SSRIs, DOACs
1
📋
Screening Check
USPSTF A/B preventive screening recommendations by age, sex, and risk factors
0
🧠
Mental Health
PHQ-9, GAD-7, AUDIT-C screening tools with severity tiers and next steps
1
Managing Your Care
Service
What happens
COIN
📊
Vital Signs
Normal ranges, concern thresholds, and critical values — BP, HR, SpO2, BMI, temperature
0
🎯
Treatment Targets
Chronic disease management targets — A1c, BP, LDL-C, eGFR by condition and guideline
1
🚨
Red Flags
Emergency recognition — ACS, stroke, sepsis, anaphylaxis, DKA, PE with time-critical actions
0
🏥
ICD-10 Lookup
Common primary care diagnosis codes for your records
1
Every answer is sourced. Every recommendation is evidenced. Every interaction is yours.
Your Journey
COIN = WORK. Every governed interaction earns COIN. General health questions are always free.
Clinical governance is structural, not advisory. MEDCHAT has your back on every question:
Your right
How MEDCHAT protects it
Evidence
Every claim cites CDC, WHO, USPSTF, AHA, ADA, NCCN, or peer-reviewed literature — no unsourced answers
No diagnosis
MEDCHAT never diagnoses or prescribes — it explains what the evidence says
Screening accuracy
Every preventive screening recommendation cites the USPSTF grade (A or B) and population criteria
Medication safety
Every drug reference includes FDA-sourced side effects, monitoring requirements, and contraindications
Emergency clarity
Red flag conditions include time-critical actions — door-to-balloon, tPA window, hour-1 bundle
Physician primacy
Every response reminds you: consult your physician for decisions
Why It Works Across Industries
MEDCHAT runs on the same engine that governs breast health at MAMMOCHAT, cancer staging at ONCOCHAT, and real estate operations at RUNNER. Same standard. Different context.
Your MEDCHAT task
Same standard as
USPSTF screening recommendation (Grade A/B)
Property appraisal (USPAP)
Chronic disease target (ADA/AHA)
Legal compliance (statute lookup)
Medication guide (FDA labeling)
Vendor credentialing (FL 468/626)
ICD-10 coding (WHO)
Financial audit (SOX compliance)
Mental health screening (PHQ-9/GAD-7)
Home inspection (FL Statute 468)
Emergency red flags (AHA/ASCO)
Post-closing coordination (CMS)
Your health navigation is governed to the same standard as a real estate transaction.
0-4 minimal; 5-9 mild; 10-14 moderate; 15-21 severe
>=10: evaluate for anxiety disorder; consider SSRI/SNRI or CBT
USPSTF
AUDIT-C
Alcohol use (3 items)
0-12
Men >=4, Women >=3: positive screen
Positive -> full AUDIT (10 items); brief intervention
USPSTF
Columbia Suicide Severity Rating Scale
Suicidal ideation and behavior
5 ideation levels + 5 behavior categories
Any active ideation with plan -> psychiatric emergency
Administer if PHQ-9 item 9 >=1
Columbia
ICD-10 Common Primary Care Codes
Code
Description
Category
Source
I10
Essential (primary) hypertension
Cardiovascular
WHO ICD-10
E11.9
Type 2 diabetes mellitus without complications
Endocrine
WHO ICD-10
E11.65
Type 2 diabetes mellitus with hyperglycemia
Endocrine
WHO ICD-10
E78.5
Dyslipidemia, unspecified
Metabolic
WHO ICD-10
J06.9
Acute upper respiratory infection, unspecified
Respiratory
WHO ICD-10
J45.909
Unspecified asthma, uncomplicated
Respiratory
WHO ICD-10
M54.5
Low back pain
Musculoskeletal
WHO ICD-10
F32.9
Major depressive disorder, single episode, unspecified
Mental health
WHO ICD-10
F41.1
Generalized anxiety disorder
Mental health
WHO ICD-10
K21.0
GERD with esophagitis
Gastrointestinal
WHO ICD-10
N39.0
Urinary tract infection, site not specified
Genitourinary
WHO ICD-10
J44.1
COPD with acute exacerbation
Respiratory
WHO ICD-10
E03.9
Hypothyroidism, unspecified
Endocrine
WHO ICD-10
G43.909
Migraine, unspecified, not intractable
Neurological
WHO ICD-10
L70.0
Acne vulgaris
Dermatologic
WHO ICD-10
COIN
Action
COIN
Evidence
Question
0
Free tier — health information access is universal
Evidence summary
1
Sourced response with guideline citations
Medication guide
1
FDA-sourced drug class reference
Screening recommendation
0
USPSTF Grade A/B — public health
Treatment target lookup
1
ADA/AHA/KDIGO guideline-sourced target
Mental health screening context
1
PHQ-9/GAD-7 severity interpretation
Persona
Field
Value
tone
warm, thorough, plain-language — CHAT never speaks without INTEL
audience
general public seeking health information, patients preparing for appointments, caregivers
voice
second-person — you are asking the right questions, here is what the evidence says
warmth
accessible guide, not a clinical manual — the governance is structural, the voice is human
context
MEDCHAT = TALK = CHAT + INTEL. Industry is primary care. Sources: CDC (100+ topics), WHO ICD-10, USPSTF, AHA, ADA, NCCN, UpToDate, DynaMed, peer-reviewed literature. Heritage: Hadley Lab clinical informatics — same engine as MAMMOCHAT, general medicine scope.
Welcome
Hey — welcome to MEDCHAT.
I am your health companion. Ask me anything about conditions, medications, screening, or wellness — every answer is evidence-based.
What can I help with today?
Service
What happens
🩺
Conditions
Evidence-based overview of common conditions
💊
Medications
Drug classes, side effects, monitoring requirements
📋
Screening
USPSTF preventive screening by age and risk factors
🧠
Mental Health
Depression and anxiety screening tools with next steps
📊
Vitals
What your numbers mean — BP, A1c, BMI, SpO2
🚨
Red Flags
When to call 911 — ACS, stroke, sepsis, anaphylaxis
Every interaction is governed. Your questions are free. Your data is yours.
Just ask. “What does an A1c of 6.8 mean?” — that is all it takes.
This is not medical advice. Always consult your physician.
Marketing Surface
Hero
Element
Value
Source
Headline
Your health questions deserve evidence, not anxiety.
Narrative — patient empowerment
Subheadline
MEDCHAT answers your health questions with evidence sourced from CDC, USPSTF, AHA, and ADA — not forum posts. Conditions, medications, screening, mental health — governed and accessible.
SHOULD cross-reference with MAMMOCHAT/ONCOCHAT for shared medical patterns
Meta-Patterns
Date
Pattern
Source
2026-02-10
ICD-10 and drug interaction patterns are the two initial learning sources
MEDCHAT LEARNING.md ledger
2026-02-14
Cross-reference constraint (MAMMOCHAT/ONCOCHAT) enables shared medical pattern propagation
MEDCHAT LEARNING.md constraint 4
2026-02-14
MEDCHAT is the general medical scope — learns from all specialist siblings
Architecture observation
2026-03
Red flags (ACS, stroke, sepsis, anaphylaxis, DKA, PE) require time-critical action windows — not just symptom lists but door-to-balloon, tPA window, hour-1 bundle
domain governance bootstrap
2026-03
USPSTF screening grades must include population criteria (age, sex, risk factors) — a Grade B without the eligible population is incomplete governance
domain governance bootstrap
2026-03
Plain language is a governance requirement not a preference — persona mandates “warm, thorough, plain-language” for general public audience
domain governance bootstrap
2026-03
Mental health screening (PHQ-9 item 9) gates to Columbia Suicide Severity Rating Scale — this escalation chain is structural, not optional
domain governance bootstrap
2026-03
Chronic disease targets differ by patient complexity — A1c <7% for most adults but <8% for elderly/complex comorbid — one-size targets are unsafe
domain governance bootstrap
2026-03-16
SESSION_LEDGERED
Explain to me aHUS
fa2400a9-d734-4cbc-be09-28dfbd98bb5b
Marketing Patterns
Date
Signal
Category
Pattern
Source
2026-03-15
COMMUNITY_BRIDGE
COMMUNITY
Community learning dashboard sourced from TALK/MEDCHAT LEARNING.md — 8 real patterns from primary care governance
Primary care navigation companion — governed health intelligence sourced from CDC, WHO, USPSTF.
USPSTF
US Preventive Services Task Force — Grade A/B recommendations for population-based screening.
CDC
Centers for Disease Control and Prevention — public health guidance and disease surveillance source.
AHA
American Heart Association — cardiovascular guidelines (BP targets, ASCVD risk, emergency protocols).
ADA
American Diabetes Association — diabetes management standards (A1c targets, SGLT2i/GLP-1 escalation).
COIN
Unit of governed work — every health navigation action earns COIN; screening recommendations are free.
INHERITANCE CHAIN
TALK
TALK is the orchestrator. CHAT is the governed view. TALK composes CHATs — each CHAT is a directed intelligence surface into the underlying scope's full INTEL graph, filtered by audience, directed by outcome. Every session governed. Every response sourced. Every question compounds community intelligence. The .ai domain is the community learning surface.
MUST: Govern every session
MUST: Source every response from evidence
MUST: Validate before deployment
MUST: Every USER principal has a dashboard at /TALKS/{USER}/
MUST: Dashboard system prompt includes identity, deals, missions, network
MUST: Dashboard system prompt declares honest capability boundaries
MUST: Every conversation turn is ledgered server-side (POST /talk/ledger)
MUST: Cross-user messages are delivered via governed inbox (POST /talk/send)
MUST: CANON.json declares users[] for cross-user message routing
MUST: Mint COIN for governed conversation work — COIN=WORK per session
MUST: Acknowledge session ledger as TRANSCRIPT-governed evidence (CHAT lane)
MUST NOT: Fabricate claims
MUST NOT: Claim capabilities the surface does not have
MUST NOT: Hardcode child scope names in law
MUST: WCAG 2.1 AA on all patient-facing surfaces (INSTANCE + COMMUNITY)
MUST: aria-live region on chat message container (screen readers announce new messages)
MUST: Skip-to-content link on all TALK layouts (bypass navigation for keyboard users)
MUST: Touch targets ≥ 44px on mobile (pointer: coarse) — clinical users include elderly patients
MUST: Focus-visible on all interactive elements (no outline:none without :focus-visible fallback)
MUST: Form controls labeled (aria-label or
TALK
TALK is CHAT + INTEL + COIN composed. Industry determines the voice. INTEL provides the knowledge. COIN gates premium access.
MUST: Wire INTEL — never speak without knowledge
MUST: Wire COIN — premium tiers require SPEND per conversation
MUST: Industry determines the voice — never generic
MUST: Every channel governed by CANON.md scope
MUST: Session ledger for every conversation turn
MUST: Every USER principal has a dashboard at /TALKS/{USER}/
MUST: Cross-user messages delivered via governed inbox
MUST: Rate-limit /chat by IP — protect provider budgets
MUST: Restrict CORS to fleet origins — never wildcard in production
MUST: Set timeouts on all external fetch calls — no unbounded waits
MUST: Log request traces with provider, latency, status
MUST NOT: TALK without disclaimer
MUST NOT: Duplicate across industries
MUST NOT: Fabricate claims
MUST: FLAGSHIP requires cross-axiomatic INTEL — sibling fleet awareness in systemPrompt
MUST: FLAGSHIP requires minimum 5 test vectors in INTEL.md
MUST: FLAGSHIP requires cross-domain routing table (when to reference legal, financial, clinical)
MUST: FLAGSHIP requires explicit COIN structure (not implicit)
MUST: FLAGSHIP requires active LEARNING ledger with SESSION_LEDGERED signals
SERVICES
SERVICES are SURFACE routes with governed boundaries.
MUST: Every service composes INTEL (mandatory — knowledge backs every operation)
MUST: Primitives are files, services are directories: INTEL→LEARNING, CHAT→TALK, COIN→SHOP
MUST: COIN is optional — adds economics (the economic shadow of WORK)
MUST: TALK is optional — adds governed conversation product
MUST: NOTIFIER is optional — adds governed cross-scope delivery (NOTIFY + INTEL)
MUST: Treat each service scope as a governed boundary (no cross-scope leakage)
MUST: Drive routes and UI from governed indices (no hardcoding)
MUST: NETWORK services MUST declare Routes table with Method + Primitive
MUST: NETWORK services MUST declare Architecture showing client → worker → backend
MUST: NETWORK services MUST declare Configuration with all vars and secrets
MUST: COMPILER services MUST declare Interface with input/output contract
MUST NOT: Embed runtime state into governed scopes
MUST NOT: Ship a service without declaring its API surface in CANON.md
SURFACE
SURFACE is the platform lane.
MUST: Keep build/sync/deploy deterministic and replay-safe
MUST: Treat GOV tree as the only source of truth
MUST: Keep runtime code minimal; enforce via governed data
MUST: Separate concerns — platform sells tiers, proof features evidence
MUST: Leaf pages with layout: paper get inheriting CANON.json — TALK walks inheritance chain
MUST: APP surface_type — Next.js App Router, reads galaxy.json + CANON.json directly
MUST: GALAXY surface_type — chrome-free operating surface (auth gate, Finder)
MUST: APP scopes compile CANON.md → CANON.json (same as SURFACE, different target)
MUST: APP scopes wire COIN + INTEL + TALK primitives into compiled CANON.json
MUST NOT: Emit new JEKYLL_* surface types — existing types frozen, deprecated
MUST NOT: Hardcode fleet membership, surfaces, or identity maps
MUST NOT: Embed proof-site URLs in platform-site content (fleet.json eco-bar only)
MAGIC
INTEL. CHAT. COIN. — Three primitives. One governed economy.
MUST: CANON.md in every scope
MUST: Services compose primitives — never duplicate
MUST: Primitive structure is fixed — industry is the only variable
MUST: Primitives compose into services — never duplicate
MUST: Services connect through SHOP.md and VAULT.md projection files
MUST: SHOP.md = public projection file (filesystem-discoverable, UPPERCASE per LANGUAGE)
MUST: VAULT.md = private projection file (filesystem-discoverable, auth-gated, UPPERCASE per LANGUAGE)
MUST: Instance = service projected through user governance context
MUST: Instance directories live at USER scope ({USER}/{PLURAL}/), not nested in SERVICES/
MUST: Service directories (SERVICES/{SINGULAR}/) define schemas — instances hold content
MUST: Every .md compiles to .json with the same name (direct mapping)
MUST: CANON.md = axiom + universal constraints only (no service names, no paths, no implementation)
MUST: README.md = how to run the CANON only
MUST: {SCOPE}.md = SPEC — the interface (purpose, routes, projections, ecosystem)
MUST NOT: Hardcode service names in CANON constraints (law speaks universals)
MUST: Inheritance resolves upward — scopes compose by directories
MUST: Tier algebra is canonical — DESIGN.md is the single source (COMPLIANCE tier algebra)
MUST NOT: Expose dimension internals to users or developers
MUST NOT: Hardcode outside governed contracts
MUST: Nonprofits get enterprise for free
MUST: ORG is the container; USER is the repo (`github.com/{org}/{user}`; duplicates across orgs allowed)
MUST: MARKET/ SALES/ GTM/ exist (META self-closure; one primitive each)
MUST: Each META sub-scope maps exactly one primitive (INTEL, CHAT, COIN)
MUST NOT: Add META business knowledge outside MAGIC/ scope
MUST NOT: Remove META sub-scope without replacing its primitive coverage
MUST: `{SCOPE}.md` is the scope contract surface; it MUST NOT be treated as a generic filename placeholder
MUST: LEARNING.md is the terminal — governance evidence, patterns, epoch rotation
MUST: LEARNING/ is the IDF directory — machine-generated individual data files
MUST: LEARNING.md rotates at epoch boundaries — frozen epochs archive as LEARNING-{EPOCH}.md at scope root
MUST: LEARNING.md is always the current epoch — active, append-only
MUST: Epoch boundary = EVOLUTION signal in LEARNING.md (named, dated, sourced)
MUST NOT: Delete archived LEARNING epochs — append-only history
MUST: MAGIC defines the triad interface directly:
MUST: COMPLIANCE/ + GALAXY/ + SURFACE/
MUST NOT: Define conflicting tier algebra in downstream scopes; downstream must inherit this contract
MEDCHAT · SERVICE CONTRACT · CANONIC ∩
🩺
MedChat
Community Learning for Medicine Residents
USPSTF screening. Chronic disease targets. Emergency red flags. Always free for students.