@csb 4 more years until AI can do it

I Built My Own iMessage Wrapped (And So Can You)

My daughter was joking that someone should build a Spotify Wrapped, but for iMessage — your year in emoji and reactions. My wife said my brother-in-law should build it since he works at Apple. (Every Apple problem is his fault in our house.)

From the other room, my daughter yelled: “I bet Dad could write this!”

Challenge accepted.

The Result

Thirty minutes later, I had a working script. Here’s my 2025:

🎁 iMessage Wrapped 2025 📱 Messages Total: 35,768 Sent: 12,191 Received: 23,577 💬 Reactions Given: 1,974 Received: 5,014 🏆 Your Reaction Style 👍 982 ❤️ 398 😂 275 ‼️ 126 ❓ 19 👎 16 📈 Messages by Month Jan ███████████████ 3,052 Feb █████████████ 2,621 Mar █████████████████ 3,422 Apr ██████████████████ 3,696 May ██████████████████ 3,640 Jun ████████████████████ 3,904 Jul █████████████ 2,561 Aug ████████████ 2,448 Sep ██████████████ 2,790 Oct █████████████████ 3,466 Nov ███████████ 2,206 Dec ██████████ 1,962

Some things I learned about myself:

I’m a listener. I receive almost 2x as many messages as I send.People love reacting to me. 5,014 reactions received vs 1,974 given.I’m a thumbs-up guy. accounts for half my reactions. Apparently I’m very agreeable.June was my chattiest month. December was quietest (holiday break mode).How It Works

Your iMessage history lives in a SQLite database at ~/Library/Messages/chat.db. It’s just sitting there, queryable.

The tricky bits:

Dates are weird. Apple stores timestamps as nanoseconds since January 1, 2001 (because of course they do).Reactions are messages. When you tapback a ️, it’s stored as a separate message with associated_message_type set to a magic number (2000 = loved, 2001 = liked, etc.).Custom emoji reactions landed in iOS 17 and are stored in associated_message_emoji.

Once you know the schema, it’s just SQL.

The Script

Here’s the full thing — about 100 lines of Python, no dependencies beyond the standard library:

#!/usr/bin/env python3 """ iMessage Wrapped — Your year in emoji and reactions Usage: python3 imessage-wrapped.py [year] Requires: Full Disk Access for Terminal """ import sqlite3 import os import sys from datetime import datetime from pathlib import Path YEAR = int(sys.argv[1]) if len(sys.argv) > 1 else 2025 DB_PATH = Path.home() / "Library/Messages/chat.db" APPLE_EPOCH_OFFSET = 978307200 TAPBACKS = { 2000: "❤️", # Loved 2001: "👍", # Liked 2002: "👎", # Disliked 2003: "😂", # Laughed 2004: "‼️", # Emphasized 2005: "❓", # Questioned } def get_db(): if not DB_PATH.exists(): print(f"❌ Database not found at {DB_PATH}") sys.exit(1) return sqlite3.connect(f"file:{DB_PATH}?mode=ro", uri=True) def date_filter(year): return f""" datetime(date/1000000000 + {APPLE_EPOCH_OFFSET}, 'unixepoch') >= '{year}-01-01' AND datetime(date/1000000000 + {APPLE_EPOCH_OFFSET}, 'unixepoch') < '{year + 1}-01-01' """ def main(): print(f"\n🎁 iMessage Wrapped {YEAR}\n") db = get_db() cur = db.cursor() # Message counts cur.execute(f""" SELECT COUNT(*), SUM(CASE WHEN is_from_me = 1 THEN 1 ELSE 0 END), SUM(CASE WHEN is_from_me = 0 THEN 1 ELSE 0 END) FROM message WHERE {date_filter(YEAR)} AND associated_message_type = 0 """) total, sent, received = cur.fetchone() print(f"📱 Messages: {total:,} ({sent:,} sent, {received:,} received)") # Reaction counts cur.execute(f""" SELECT SUM(CASE WHEN is_from_me = 1 THEN 1 ELSE 0 END), SUM(CASE WHEN is_from_me = 0 THEN 1 ELSE 0 END) FROM message WHERE {date_filter(YEAR)} AND associated_message_type >= 2000 """) given, got = cur.fetchone() print(f"💬 Reactions: {given + got:,} ({given:,} given, {got:,} received)") # Your tapback style print(f"\n🏆 Your Reaction Style") cur.execute(f""" SELECT associated_message_type, COUNT(*) FROM message WHERE {date_filter(YEAR)} AND associated_message_type BETWEEN 2000 AND 2005 AND is_from_me = 1 GROUP BY associated_message_type ORDER BY COUNT(*) DESC """) for type_id, cnt in cur.fetchall(): print(f" {TAPBACKS.get(type_id, '?')} {cnt:,}") # Custom emoji cur.execute(f""" SELECT associated_message_emoji, COUNT(*) FROM message WHERE {date_filter(YEAR)} AND associated_message_emoji IS NOT NULL AND is_from_me = 1 GROUP BY associated_message_emoji ORDER BY COUNT(*) DESC LIMIT 5 """) customs = cur.fetchall() if customs: print(f"\n🎯 Custom Reactions: {', '.join(f'{e} ({c})' for e, c in customs)}") # Monthly volume print(f"\n📈 By Month") cur.execute(f""" SELECT strftime('%m', datetime(date/1000000000 + {APPLE_EPOCH_OFFSET}, 'unixepoch')), COUNT(*) FROM message WHERE {date_filter(YEAR)} AND associated_message_type = 0 GROUP BY 1 ORDER BY 1 """) months = ["Jan","Feb","Mar","Apr","May","Jun","Jul","Aug","Sep","Oct","Nov","Dec"] data = {row[0]: row[1] for row in cur.fetchall()} max_cnt = max(data.values()) if data else 1 for i, name in enumerate(months, 1): cnt = data.get(f"{i:02d}", 0) bar = "█" * int(20 * cnt / max_cnt) print(f" {name} {bar} {cnt:,}") db.close() if __name__ == "__main__": main()Running ItSave the script as imessage-wrapped.pyGrant Full Disk Access to your terminal (System Settings → Privacy & Security → Full Disk Access)Run it:python3 imessage-wrapped.py # defaults to 2025 python3 imessage-wrapped.py 2024 # or any year

That’s it. Your data never leaves your machine.

What I’d Add Next

If I turn this into a proper app:

Shareable cards — export your stats as an imageConversation breakdown — who do you text the most?Time of day patterns — are you a morning texter or a midnight scroller?Streak tracking — longest daily conversation streak

But honestly? The script is fun enough. Sometimes a quick hack that makes your daughter laugh is the whole point.

The script and a slightly more polished version are on GitHub if you want to grab it.

#Apple #iMessage #Messages #NerdyStuff #Python

@dave

Nope, for coding I use OpenAI codex

Openclaw I use for reminders (cron jobs) and skills (just testing yet)

@csb For coding agent work?