HUNT READY 30 — Full Build
A faith-rooted, hunting-specific fitness PWA built for short home workouts with adjustable dumbbells and a weight vest / ruck.
Current feature set
- 8-week, 30-min-or-less training program with dumbbell and bodyweight (+ bands) equipment modes
- Per-set weight and rep logging
- Carry, plank, ruck, circuit, and timed-work logging
- Automatic current working-weight updates
- Exercise progression charts
- Detailed workout history
- Hunter-readiness benchmarks
- Body-weight and waist tracking
- Daily subjective readiness check-in
- Daily Scripture and training focus
- Exercise library with animated form drawings, cues, mistakes, and inline demo videos
- Archery Lab: practice log with AI target measurement, trends by distance, AI form check from a short video, and bow photo scan
- Hunt-prep gear / legal checklist
- Nutrition targets and manual meal tracker
- Camera/photo meal journal
- Optional AI meal-photo analysis through a secure server endpoint
- Offline PWA support
- Complete JSON backup / restore, including stored meal photos
Important: AI meal-photo analysis
This is a static GitHub Pages app. Do not put a private Anthropic (Claude) or other AI provider API key directly into app.js, browser localStorage, or the Settings field.
The app expects an optional secure server endpoint configured in Settings. It POSTs:
{
"imageDataUrl": "data:image/jpeg;base64,..."
}
The endpoint should return JSON like:
{
"name": "Steak and eggs",
"calories": 820,
"protein": 72,
"carbs": 4,
"fat": 58,
"fiber": 0,
"notes": "Estimated from photo; review portion sizes."
}
Photo-derived nutrition should always be treated as an estimate and reviewed before saving.
GitHub Pages update
Replace these files when publishing a new version:
index.htmlapp.jsstyles.csssw.jsmanifest.webmanifest
The icons folder only needs replacing when the app icon changes.
AI backend on Vercel (Claude via Vercel AI Gateway)
api/analyze-meal.js is a Vercel serverless function that sends the meal photo to Claude using the official @anthropic-ai/sdk with structured JSON output. By default it routes through Vercel AI Gateway, so the only credential needed is an AI Gateway key created inside the Vercel dashboard (AI Gateway → API Keys). The gateway includes $5/month of free credits once a credit card is on file and charges provider list price with no markup after that.
Required Vercel environment variables:
AI_GATEWAY_API_KEY — Secret. Your Vercel AI Gateway key (starts with vck_). If the same key is stored as ANTHROPIC_API_KEY instead, it is detected automatically.ALLOWED_ORIGINS — comma-separated allowed browser origins, for example https://fabiano-create.github.ioANTHROPIC_MODEL — optional; defaults to anthropic/claude-opus-5
Alternative: set ANTHROPIC_API_KEY to a direct Anthropic key (sk-ant-…) from console.anthropic.com and the function calls Anthropic directly instead.
After adding or changing variables, redeploy the latest production deployment so the function picks them up.
The endpoint URL (https://hunt-ready-30.vercel.app/api/analyze-meal) goes into HUNT READY 30 → Settings → Secure AI nutrition endpoint. The archery features use api/analyze-archery.js at the same host; the app derives that URL automatically.
Error responses are JSON { "error": "...", "code": "..." } and the app shows the message directly. HTTP 402 means the connected AI account has no credits.
