Daily Recap

Daily Recap is your private end-of-day read. It gathers everything you did with AI today into one page: a grade for the day, how it compares to yesterday, a plain-language account of what you worked on, and the numbers behind it all. It counts only sub-sessions that have closed, and only you ever see it.

Today’s verdict

The header card gives you the day in one look: today’s letter grade and Prompt Score, the change against yesterday (▲ / ▼), and a short written verdict — a sentence or two on how the day went. Below it sit three vs-yesterday tiles, each with today’s value and the change since yesterday:

• Completion rate — the share of today’s sub-sessions that you crushed.
• Retries — how much of today’s work had to be redone as rework; lower is better.
• Sessions — how many Claude Code sessions you ran.

What you did today

• Today’s work — a written narrative of what you actually worked on, so you don’t have to reconstruct the day from raw numbers.
• Model usage — how the day split across models, as a share bar and per-model list, with your most- and least-used called out.
• Files edited — the directories you touched most, as a bar chart.
• Top tasks today — a table of your biggest tasks, with Task, Time, Sessions, and Tokens.

Today’s AI-use verdict

Four cards give you the day’s coaching — and this section is visible only to you, never to a manager:

• Best / worst prompt — your most-missed prompt of the day, with an appeal form if you disagree with the call.
• Today — quantitative — the hard numbers (completion, retries, sessions).
• Today — qualitative — a written, AI-authored read on how you worked.
• Today — average Prompt Grade — your mean prompt grade for the day.

Standing and appeals

• Where you stand vs your team — a card that links to your Prompt Score analytics, where your percentile and strengths live.
• My appeals — the grade disputes you’ve filed, with their status. When you appeal the “worst prompt” card, it shows up here so you can track it.