Luna AI Tools Reference

Luna has 13 specialized tools organized into 7 categories. Each tool is a function that Luna can call during a conversation to retrieve data, perform analysis, or record information.

Health Data Tools

checkContraindication

Check if a peptide or supplement is contraindicated for specific medical conditions.

Parameters:

Example interaction:

User: “Is BPC-157 safe for someone with type 2 diabetes?”

Luna calls: checkContraindication({ productId: "bpc-157", conditions: ["type-2-diabetes"] })

What it does:

1. Resolves product synonyms via resolveSynonym('peptides', productId)
2. Fetches contraindication data from getContraindicationsForPeptide()
3. Matches user conditions against known contraindications
4. Returns allowed/blocked status with severity and reasoning

Returns:

{
  "allowed": true,
  "peptide": "bpc-157",
  "matches": [],
  "warnings": ["Monitor blood sugar levels closely"]
}

getBiomarkerRange

Get normal and optimal reference ranges for a specific biomarker.

Parameters:

Example interaction:

User: “What’s a normal testosterone level for a 35-year-old male?”

Luna calls: getBiomarkerRange({ biomarkerId: "testosterone-total", age: 35, sex: "male" })

Returns:

{
  "biomarker": "testosterone-total",
  "name": "Total Testosterone",
  "unit": "ng/dL",
  "ranges": {
    "low": 300,
    "high": 1000,
    "optimalLow": 500,
    "optimalHigh": 900
  },
  "interpretation": "Optimal range for males aged 30-39: 500-900 ng/dL"
}

getRequiredPanels

Get required lab panels for a specific protocol or peptide.

Parameters:

Example interaction:

User: “What labs do I need before starting ipamorelin?”

Luna calls: getRequiredPanels({ protocolId: "ipamorelin" })

Returns:

{
  "protocolId": "ipamorelin",
  "panels": [
    {
      "id": "gh-panel",
      "name": "Growth Hormone Panel",
      "biomarkers": [
        { "code": "igf-1", "name": "IGF-1", "required": true },
        { "code": "gh", "name": "Growth Hormone", "required": false }
      ]
    },
    {
      "id": "comprehensive-metabolic",
      "name": "Comprehensive Metabolic Panel",
      "biomarkers": [...]
    }
  ]
}

searchBiomarkers

Search biomarkers by name, synonym, or description.

Parameters:

Example interaction:

User: “What biomarker measures thyroid function?”

Luna calls: searchBiomarkers({ query: "thyroid" })

Returns:

{
  "results": [
    { "code": "tsh", "name": "TSH", "category": "thyroid" },
    { "code": "free-t3", "name": "Free T3", "category": "thyroid" },
    { "code": "free-t4", "name": "Free T4", "category": "thyroid" }
  ]
}

Patient Graph Tools

getPatientContext

Fetch the complete patient context including profile, conditions, medications, labs, and protocols.

Parameters:

What it returns:

• Patient profile (demographics, subscription tier)
• Active conditions
• Current medications
• Recent lab results
• Active protocols
• Emergency flags (if any)
• Consultation notes

Example interaction:

User: “Review my health profile and suggest improvements.”

Luna calls: getPatientContext({ userId: "user_clerk_123" })

checkDrugInteractions

Check potential interactions between a proposed medication and the patient’s current medications.

Parameters:

What it does:

1. Loads patient context to get current medications
2. Checks against a built-in interaction database
3. Generates interaction warnings with severity levels

Returns:

{
  "proposedMedication": "BPC-157",
  "currentMedications": ["warfarin", "metformin"],
  "interactions": [
    {
      "medication": "warfarin",
      "severity": "moderate",
      "description": "BPC-157 may affect platelet aggregation; use with caution alongside anticoagulants"
    }
  ],
  "safeToUse": true,
  "warnings": ["Monitor INR levels more frequently"]
}

getLabResults

Get recent lab results with optional biomarker filtering and trend analysis.

Parameters:

What it does:

1. Fetches lab results from the Patient Graph
2. Optionally filters by biomarker code
3. Computes trends via analyzeBiomarkerTrends()

Example interaction:

User: “How have my testosterone levels changed over time?”

Luna calls: getLabResults({ userId: "user_clerk_123", biomarkerCode: "testosterone-total" })

getActiveProtocols

Get all active protocols for a patient with compound details.

Parameters:

Returns:

{
  "protocols": [
    {
      "id": "proto_abc",
      "title": "BPC-157 Recovery Protocol",
      "status": "active",
      "startDate": "2024-06-01",
      "items": [
        {
          "compound": "BPC-157",
          "dosage": "250mcg",
          "frequency": "2x daily",
          "route": "subcutaneous"
        }
      ]
    }
  ]
}

Protocol Tools

trackProtocol

Log a protocol adherence check-in and calculate the adherence streak.

Parameters:

What it does:

1. Creates a note_added event in the Patient Graph
2. Queries last 90 days of events to compute streak
3. Returns current streak count and motivational feedback

Example interaction:

User: “Just took my BPC-157 this morning”

Luna calls: trackProtocol({ protocolName: "BPC-157", notes: "Morning dose taken" })

Luna responds: “Nice! That’s 14 days in a row. Keep it up!”

Protocol Analysis Tools

analyzeProtocol

Analyze protocol adherence trends and biomarker changes to suggest adjustments.

Parameters:

What it does:

1. Gathers check-in data for the analysis period
2. Collects biomarker trends during the protocol
3. Sends data to GPT-4-turbo for clinical analysis
4. Returns adherence summary, biomarker changes, and AI-generated suggestions

Returns:

{
  "adherence": {
    "rate": 0.85,
    "missedDays": 7,
    "totalDays": 48,
    "streak": 14
  },
  "biomarkerTrends": [
    {
      "code": "igf-1",
      "startValue": 180,
      "currentValue": 245,
      "change": "+36%",
      "trend": "improving"
    }
  ],
  "suggestions": [
    "Consider maintaining current dosage — IGF-1 levels are responding well",
    "Address the missed doses on weekends with reminders"
  ]
}

Lab Analysis Tools

analyzeLabReport

Parse and interpret an uploaded lab report PDF.

Parameters:

What it does:

1. Downloads or decodes the lab report PDF
2. Uses @loop/biomarker-parser with Anthropic to extract biomarkers
3. Matches raw names to canonical codes
4. Normalizes units to standard units
5. Classifies values against sex-specific reference ranges
6. Records a lab_parsed event in the Patient Graph
7. Returns interpreted results with reference ranges

Example interaction:

User: (uploads lab report PDF)

Luna calls: analyzeLabReport({ fileUrl: "data:application/pdf;base64,...", userId: "user_123", patientSex: "male" })

Luna responds: “I’ve analyzed your lab report. Here are the key findings: Your testosterone is at 650 ng/dL (optimal range), your TSH is slightly elevated at 4.2 mIU/L…”

Wearable Insights Tools

getWearableInsights

Get wearable data insights with trend analysis and optional protocol correlation.

Parameters:

What it does:

1. Queries wearableData repository for the specified metrics and date range
2. Computes summaries (averages, min, max) for each metric type
3. Detects trends (improving, stable, declining)
4. Generates metric-specific recommendations
5. Optionally correlates with active protocols

Returns:

{
  "metrics": {
    "sleep": {
      "average": 7.5,
      "trend": "improving",
      "dataPoints": 14,
      "summary": "Sleep quality has improved 12% over the past 2 weeks",
      "recommendations": ["Maintain current sleep schedule"]
    },
    "hrv": {
      "average": 48,
      "trend": "stable",
      "dataPoints": 14,
      "summary": "HRV is stable at 48ms",
      "recommendations": ["Consider stress management techniques"]
    }
  },
  "protocolCorrelation": {
    "protocol": "BPC-157 Recovery",
    "finding": "Recovery scores improved 15% since protocol start"
  }
}

Check-In Tools

analyzeCheckIn

Analyze weekly check-in data, celebrate progress, and troubleshoot issues.

Parameters:

Check-In Data Schema:

What it does:

1. Compares current metrics with previous check-in
2. Identifies wins (improvements of 2+ points)
3. Flags concerns (declines of 2+ points)
4. Notes side effects if present
5. Adds protocol-phase-specific suggestions
6. Builds a coaching-style response

Example interaction:

User: “Weekly check-in: pain 3, sleep 8, mood 7, energy 6”

Luna calls: analyzeCheckIn({ currentCheckIn: { pain_level: 3, sleep_quality: 8, mood: 7, energy: 6 }, protocolName: "BPC-157", week: 4 })

Luna responds: “Great progress this week! Your sleep quality improved from 6 to 8 — that’s a significant jump. Pain is down to 3 which is right on track for week 4 of BPC-157…”