Huckleberry Baby Tracker

Track baby activities (sleep, feeding, diapers, growth) via the Huckleberry app's Firebase backend.

Setup

1. Install the API:

   # Install from GitHub (required for bottle feeding support until next PyPI release)
   pip install git+https://github.com/Woyken/py-huckleberry-api.git
   # or with uv:
   uv pip install git+https://github.com/Woyken/py-huckleberry-api.git
   

2. Configure credentials (choose one):

   • Environment variables:

     export HUCKLEBERRY_EMAIL="your-email@example.com"
     export HUCKLEBERRY_PASSWORD="your-password"
     export HUCKLEBERRY_TIMEZONE="America/Los_Angeles"  # optional
     

   • Config file at ~/.config/huckleberry/credentials.json:

     {
       "email": "your-email@example.com",
       "password": "your-password",
       "timezone": "America/Los_Angeles"
     }
     

CLI Usage

The CLI is at ~/clawd/skills/huckleberry/scripts/hb.py

# List children
python3 ~/clawd/skills/huckleberry/scripts/hb.py children

# Sleep tracking
python3 ~/clawd/skills/huckleberry/scripts/hb.py sleep-start
python3 ~/clawd/skills/huckleberry/scripts/hb.py sleep-pause
python3 ~/clawd/skills/huckleberry/scripts/hb.py sleep-resume
python3 ~/clawd/skills/huckleberry/scripts/hb.py sleep-complete
python3 ~/clawd/skills/huckleberry/scripts/hb.py sleep-cancel

# Breastfeeding
python3 ~/clawd/skills/huckleberry/scripts/hb.py feed-start --side left
python3 ~/clawd/skills/huckleberry/scripts/hb.py feed-switch
python3 ~/clawd/skills/huckleberry/scripts/hb.py feed-pause
python3 ~/clawd/skills/huckleberry/scripts/hb.py feed-resume --side right
python3 ~/clawd/skills/huckleberry/scripts/hb.py feed-complete
python3 ~/clawd/skills/huckleberry/scripts/hb.py feed-cancel

# Bottle feeding
python3 ~/clawd/skills/huckleberry/scripts/hb.py bottle 120 --type "Formula" --units ml

# Diaper
python3 ~/clawd/skills/huckleberry/scripts/hb.py diaper both --pee-amount medium --poo-amount big --color yellow --consistency loose

# Growth
python3 ~/clawd/skills/huckleberry/scripts/hb.py growth --weight 5.2 --height 55 --head 38 --units metric
python3 ~/clawd/skills/huckleberry/scripts/hb.py growth-get

# History
python3 ~/clawd/skills/huckleberry/scripts/hb.py history --date 2026-01-27
python3 ~/clawd/skills/huckleberry/scripts/hb.py history --days 7 --type sleep --type feed
python3 ~/clawd/skills/huckleberry/scripts/hb.py history --json

Complete Parameter Reference

Sleep Commands

Breastfeeding Commands

Bottle Feeding

bottle <amount> [options]

Diaper Change

diaper <mode> [options]

Color Guide

• yellow — Normal for breastfed babies
• brown — Normal for formula-fed or older babies
• green — Can be normal, or indicate fast digestion/foremilk
• black — Normal first few days (meconium), concerning later
• red — May indicate blood, consult pediatrician
• gray — Uncommon, may indicate liver issues

Consistency Guide

• solid — Formed stool
• loose — Soft but not watery
• runny — Watery consistency
• mucousy — Contains mucus
• hard — Firm, may indicate constipation
• pebbles — Small hard pieces
• diarrhea — Very watery, frequent

Growth Measurements

growth [options]
growth-get

History / Calendar

history [options]

Use --type multiple times to filter: --type sleep --type feed

Agent Guidelines: When to Ask for Details

AI Attribution on Notes

Always include AI attribution when logging entries:

Creating new entries:

• No user note: --notes "Created via AI"
• User provides note: --notes "user's note | Created via AI"

Editing existing entries:

• No user note: --notes "Updated via AI"
• User provides note: append | Updated via AI to existing notes

This creates a paper trail for AI-assisted entries.

When to Ask for Clarification

When a user request is sparse, ask for clarification before logging. Here's when:

Diaper Changes

If user says just "diaper change" or "poop":

• Always ask: Was it pee, poo, or both?
• For poo, consider asking: Color? Consistency? Amount?
• Skip details if: User seems rushed or says "just log it"

Example follow-up:

"Got it! Was it pee, poo, or both? Any details to note (color, consistency, amount)?"

Bottle Feeding

If user says "bottle" without amount:

• Always ask: How much? (in ml or oz)
• Consider asking: Formula, breast milk, or mixed?

Example follow-up:

"How much was the bottle? And was it formula, breast milk, or mixed?"

Growth Measurements

If user says "log weight" without value:

• Always ask: What's the weight? (and clarify units if ambiguous)

Sleep/Feeding Timers

These are typically clear commands, but clarify if ambiguous:

• "Baby's eating" → "Starting breastfeeding — which side, left or right?"
• "Feed done" → Check if breastfeeding or bottle context

Natural Language Examples

Multi-Child Support

If the account has multiple children, use --child / -c:

python3 hb.py --child "Baby Name" sleep-start

Without --child, commands default to the first child in the account.

Troubleshooting

Authentication errors:

• Verify email/password are correct
• Check credentials file permissions
• Huckleberry doesn't support 2FA for API access

"No children found":

• Ensure the account has at least one child profile in the Huckleberry app

Timer already active:

• Complete or cancel the existing session before starting a new one

Technical Notes

• Uses Firebase Firestore via gRPC (same as mobile app)
• Real-time sync: Changes appear immediately in the Huckleberry app
• Token auto-refresh: Sessions stay authenticated
• Timezone handling: Huckleberry requires an offset field (minutes behind UTC) for entries to display correctly. E.g., PST (UTC-8) = 480 minutes. The CLI automatically calculates this from the configured timezone. Without this field, entries appear at UTC time in the app.

---

Credits

Built on py-huckleberry-api by Woyken — a reverse-engineered Python client for Huckleberry's Firebase backend.

---

Created with AI - 2026-01-27 Updated with AI - 2026-01-28

Notes on All Entry Types

The --notes / -n parameter is available on all entry types:

• sleep-complete --notes "Slept through the night!"
• feed-complete --notes "Good latch today"
• bottle --notes "Logged via AI"
• diaper --notes "Checked by AI"
• growth --notes "Measured at pediatrician"

The upstream py-huckleberry-api only supports notes on diaper entries. This skill extends that to all types by updating the Firestore document directly after creation.

Not Supported by Upstream API

The following features exist in Huckleberry but aren't exposed in the py-huckleberry-api:

• Sleep conditions (happy/upset at start/end)
• Sleep locations (car, nursing, stroller, crib, etc.)

These would require modifying the upstream library to accept additional parameters.