Skip to main content
A lab report tells a user what is out of range. This flow turns it into a plan: each out-of-range finding maps to lifestyle interventions and an evidence-graded supplement stack, personalized against what the user already takes, with drug interactions flagged and cited from the literature.
StepEndpoint
Upload labsPOST /imports/file (biomarkers)
Upload supplements & medsPOST /imports/file (behavioral)
AnalyzePOST /biomarkers/analyze
Build the protocolGET /analyses/{id}/action-plan

​
1. Upload labs and the user’s current stack

LABS=$(curl -s -X POST "$FB_API/imports/file" -H "authorization: Bearer $FB_KEY" \
  -H "content-type: application/json" -d '{
    "user_id":"'$UID'","organization_id":"'$ORG'","category":"biomarkers",
    "filename":"labs.csv","content_type":"text/csv",
    "text":"marker,value,unit\nApoB,132,mg/dL\nTriglycerides,190,mg/dL\nVitamin D,21,ng/mL\nHomocysteine,14,umol/L"
  }' | jq -r .source.id)

# Behavioral upload records what they already take, used to personalize the plan.
BEH=$(curl -s -X POST "$FB_API/imports/file" -H "authorization: Bearer $FB_KEY" \
  -H "content-type: application/json" -d '{
    "user_id":"'$UID'","organization_id":"'$ORG'","category":"behavioral",
    "filename":"stack.json","content_type":"application/json",
    "text":"{\"entries\":[{\"kind\":\"supplement\",\"name\":\"Omega-3\",\"dose\":\"2 g\"},{\"kind\":\"medication\",\"name\":\"Warfarin\",\"dose\":\"5 mg\"}]}"
  }' | jq -r .source.id)

​
2. Analyze

AN=$(curl -s -X POST "$FB_API/analyses" -H "authorization: Bearer $FB_KEY" \
  -H "content-type: application/json" \
  -d '{"user_id":"'$UID'","organization_id":"'$ORG'","source_ids":["'$LABS'","'$BEH'"],"profile":{"age":45,"sex":"male"}}' \
  | jq -r .id)

​
3. Build the action protocol

curl -s "$FB_API/analyses/$AN/action-plan" -H "authorization: Bearer $FB_KEY" | jq .
What comes back:
  • Interventions first: lifestyle changes (reduce saturated fat, zone-2 cardio) ranked core vs optimize, each with the markers it targets.
  • An evidence-graded supplement discussion list: A–D grade, mechanism, and which flagged markers it addresses. Exact dose and timing are withheld by default until a clinician-reviewed deployment policy explicitly enables them.
  • Personalized to their stack: supplements they already log are marked already_taking instead of re-recommended.
  • Cited interaction cautions: because the user takes warfarin, the plan flags it: “Vitamin K2 has a documented interaction with Warfarin (186 literature reports via supp.ai).” Interaction cautions only appear for drugs the user actually logs.
{
  "supplements": [{
    "id": "omega_3", "name": "Omega-3 (EPA/DHA)",
    "dose_guidance": "Exact dose and timing are withheld in the wellness API.",
    "evidence": "A", "priority": "core",
    "targets": [{"marker":"triglycerides","direction":"high"},{"marker":"apob","direction":"high"}],
    "already_taking": true,
    "cautions": ["You logged Warfarin: Omega-3 can increase bleeding risk…"],
    "sources": [{"name":"SUPP.AI","url":"https://supp.ai/…"},{"name":"Pillser","url":"https://pillser.com/…"}]
  }],
  "disclaimer": "Educational wellness guidance, not medical advice…"
}

​
Take it further

  • Add genetics to close the pharmacogenomics loop: CPIC gene-drug context surfaces at the plan level.
  • Re-run after the next panel to watch the protocol evolve.
  • Always surface the disclaimer and sources in your UI.