Internal, privacy-sensitive tool. Tracks route/method/status/timing + account, not page content or request bodies — built to be pruned or removed from production if it's judged too invasive. Loading status…
👣 Visitors & engagement
Everything below "Traffic overview" starts at an API call from someone signed in. This is the part before that: who arrives, what they open first, how long they stay, and whether they come back.
Opened first — what actually drew them in
Where they came from — referring host or campaign tag
Modules used — everywhere a visit went
Do accounts come back? — visits per signed-in user
What returners use — signed-in visits only
📉 Traffic overview
🧩 Usage by module
Requests, active users and error rate per product module over the selected window — click a module to drill into its routes.
Routes
Route
Method
Requests
Errors
Active users (approx)
Avg latency
🧊 Cold features
Modules with little or no traffic over the window — candidates to review, simplify, or cut.
🧩 Puzzle solving & return visits
Who actually solves puzzles, and whether solving brings them back. The second half is the one that matters: a solve that ends a session and a solve that starts a habit look identical in an attempt count. Return is measured per user on their most recent solve, so one prolific solver cannot carry the rate.
Loading…
📧 Daily puzzle email
Who is sent the daily puzzle email, who opens the puzzle from it, and who solves. “Opened” counts the email’s own link, signed for that person and that day; mail scanners that open every link the moment a message arrives are left out. “Solved” is that day’s daily puzzle, or any puzzle that day.
Loading…
🗓 Structure review schedule
The 1/3/7/16/35/90-day ladder is the conventional shape, borrowed rather than measured on this audience. Two things say whether it fits: do people come back when a review falls due, and does accuracy hold as the gaps widen? A rung where answers suddenly go wrong is a rung that came too late — the interval before it is the number to change.
💳 Usage by subscription tier
Which modules each tier actually touches — does a gated feature justify its cost?
🐛 Error hotspots
Routes with the most errors across every module over the window.
Module
Route
Method
Requests
Errors
Error rate
🗄 Log maintenance
Every log that can be cleared, with what it currently costs. Cleared one at a time on purpose — the reason to clear request logs (live sizes in the table below) is not the reason to clear a security record, and one button for both invites a mistake nobody can undo.
ChessMoves AI