
Coding Wrapped
See how you really vibe code with Claude Code & Codex
70 followers
See how you really vibe code with Claude Code & Codex
70 followers
Your agent remembers every session. You don't. Coding Wrapped turns your local Claude Code & Codex history into a pixel-art Wrapped: your rhythm, the words you repeat ("continue", anyone?), what the agent quietly took over, and one small tip for each habit. Open source, one command, 100% local.







Free
Launch Team / Built With

AWS Startup AdvisorPersonalized AWS guidance that meets you where you’re already building
Promoted


Hi Product Hunt, I'm Sen, a designer who never formally learned to code.
Over the past year I've built most of my projects by talking to Claude Code and Codex. Things got made, but I rarely saw how I actually worked with the agent: how I give direction, when I correct it, when I just let it continue.
Coding Wrapped is a small open-source Agent Skill, loosely inspired by Spotify Wrapped. It reads your local Claude Code and Codex sessions and turns them into a private dashboard on localhost:
A few facts about your habits. For me: I used "first" in 29 messages this month, and said "continue" 14 times once a direction felt right.
What you did, what the agent did, and the working style that shows.
One small suggestion per habit, drawn from a curated set of 40 practices, each with its source.
Everything stays on your machine. Raw prompts, code, project names and secrets are not stored, and it's not a productivity score.
To try it:
Then ask your agent: "Build my Coding Wrapped."
I'd be curious to hear what it shows you.
Mine would be "go" by a mile, that's how I answer my agents, go 1, 3 and so on. Does it work on sessions in French? I type to Claude in French most of the time
congrats on the launch sen. the words you repeat part is funny and a bit scary ) id love to see which prompts made the agent redo the same work, thats where i lose most of my time
Fun concept! Local data is great. Questions: handles multi-project setups? Export team insights? Other LLM integrations? Real-time mode beyond session review?
The local-only angle is what makes this click for me. I like that it turns repeated words like “continue” into something concrete, and that each habit gets just one sourced suggestion from those 40 practices instead of another dashboard to interpret. Curious whether the repeated-word pattern looks very different between Claude Code and Codex sessions for you?