VibeCheck — iMessage Analyzer

What I Built

A command-line tool that reads your iMessage database, analyzes conversation patterns, and gives you insights about your texting relationships — who you text most, sentiment analysis, conversation dynamics.

Tools Used

• -Claude Code — AI coding agent in the terminal
• -Python — Core language
• -Rich — Beautiful terminal output
• -SQLite — Reading the iMessage database directly

How I Prompted

Step 1: Start with the core idea

Build a Python CLI that reads my Mac's iMessage database
(~/Library/Messages/chat.db) and analyzes my conversations.
Show me who I text most, message frequency, and basic sentiment.
Use Rich for pretty terminal output.

Claude Code scaffolded the entire project — database reader, analysis engine, and Rich-formatted output.

Step 2: Add AI analysis

Now add a feature that uses Claude API to analyze the vibe
of a conversation. Pick the last 50 messages with a contact
and give me a "vibe check" — are we close? drifting? formal?

This added the AI layer — it reads recent messages and generates a relationship summary.

Step 3: Polish the output

Make the output more visual. Add emoji indicators for vibe,
a bar chart for message frequency by day of week, and color
code the sentiment scores.

Final polish made it feel like a real product.

What I Learned

• -Start broad, refine narrow. The first prompt got 80% of the functionality. The next two refined it.
• -Let the AI read real data. Pointing Claude Code at the actual iMessage database meant it could figure out the schema itself.
• -Rich makes CLIs beautiful. Zero web skills needed — terminal output can look amazing.

Try It Yourself

1. 1.Install Claude Code: npm install -g @anthropic-ai/claude-code
2. 2.Start with: "Build a Python CLI that reads [some data source] and shows [some insight]"
3. 3.Let it scaffold, then iterate with refinement prompts