💡 Voice-commanded treatment calculator with real-time dosage in zero-tap mode

[github-actions[bot]]

Summary

During a zero-tap voice-driven inspection, a beekeeper can say "treat this hive with oxalic acid dribble" and receive instant spoken dosage calculation based on hive configuration, estimated cluster size, ambient temperature, and treatment history — all without touching the phone. The system cites the product label, regional extension guidelines, and logs the treatment automatically.

Market Signal

VarroaVault ($29/month) has validated that AI-assisted treatment recommendations with automatic efficacy scoring from paired mite counts is a paid-tier feature. HiveLog AI provides zone-aware treatment schedules with dosage tables and confidence scores. Both require manual data entry — no competitor delivers treatment calculation via voice during hands-busy field work. The treatment management category is maturing rapidly as table stakes for serious beekeeping apps.

Sources: VarroaVault, HiveLog AI, Best Beekeeping Apps 2026

User Signal

Existing Discussion #85 (Treatment compliance, PHI alerts & efficacy scoring) and #348 (Automated treatment efficacy scoring) address treatment tracking and post-treatment analysis. This proposal addresses the upstream decision moment: real-time dosage calculation at the point of action, delivered voice-first within the zero-tap inspection session. PRD compliance requirements specify treatment recommendations must include "verify treatment suitability for your specific conditions" disclaimers.

Technical Opportunity

Architecture already includes Gemini STT/TTS via Vertex AI (Epic 11), voice-first inspection flow, and treatment/recommendation schema (Epic 3 Story 3.3). Treatment dosage rules are well-documented by university extensions and EPA labels — this is algorithmic, not ML. The recommendation contract (action + rationale + confidence + fallback) naturally wraps dosage output. Domain data is structured and auditable: oxalic acid concentration × cluster size × temperature range = dosage.

Assessment

Dimension	Score	Rationale
Feasibility	high	Dosage rules are algorithmic; STT/TTS pipeline exists; treatment schema is planned
Impact	high	Directly reinforces zero-tap differentiator at a high-value decision moment
Urgency	med	Depends on voice pipeline (Epic 11) being operational first

Adversarial Review

Strongest objection: Dosage liability risk — if the calculator recommends the wrong amount and colonies are harmed, user trust is permanently destroyed. Regulatory variations across US states and EU countries add complexity.

Rebuttal: Frame as "calculator" not "prescription" — show the math, cite the source (EPA labels, university extension guidelines), include persistent disclaimer per PRD compliance requirements. VarroaVault already delivers similar functionality without reported liability issues. The value is convenience at the moment of action, not medical authority. Treatment databases are well-structured, versioned, and auditable by jurisdiction.

Suggested Next Step

Compile treatment dosage rules for the top 5 treatments (oxalic acid dribble/vaporization, formic acid, ApiVar, HopGuard, Apivar) across US/EU jurisdictions as a domain data artifact in _bmad-output/. Design the voice interaction flow ("treat [hive] with [treatment]" → dosage response → confirmation → auto-log) within the existing zero-tap session design.