One line per session: the day, what you worked on, the tempo or pass rate you actually reached, and an honest sentence about whether it moved. Then a read on the last 7 and 30 days — which stays silent about any focus that has not earned a verdict yet.
The problem this fixes
Most practice records answer the easy question. They count days, minutes, and sessions — all things you can produce without playing anything better at the end of them. A month of that leaves you with a number that went up and a passage that did not.
The hard question is whether anything moved, and it needs two things a streak never asks for: a specific focus, and a number attached to it. Log those, and after three sessions on the same focus the comparison starts to mean something. Log them for a month and you can tell which of the things you changed was the thing that worked. That is the same loop the GuitarHub method runs on, and the reason deliberate practice asks for a measurement rather than a duration.
What the log can support
There is no consecutive-day counter here, and there is not going to be. A streak goes up on a session that changed nothing and resets on a week off that changed a lot. These numbers only move when the playing does.
Nothing logged yet. Log 3 sessions on the same focus and this turns into a direction — before that it stays a list, because 2 points would only draw a line through a good day and a bad one.
No account and no upload. Export writes a plain JSON file you can read in any text editor, keep as a backup, or open on another machine. Importing adds those sessions to this browser — anything identical is skipped rather than duplicated.
How this works
A practice tool earns its answers by being willing to give none. These six refusals are enforced in the code that computes the summary, not in the copy describing it.
Deliberately absent, and it is not an oversight. A streak measures attendance: it goes up on a session that changed nothing, and it drops to zero on a week off that changed a lot. It is a number that always moves, which is exactly why it feels like evidence and is not.
Two points draw a straight line through any pair of numbers, including a good day next to a bad one. Under three sessions on a focus, the tool returns not enough evidence yet. That rule lives in the code rather than in this paragraph, so nothing built on top of it can talk its way past it.
If your average went up but your worst session went down with it, that is a wider spread, not progress, and the log says so. The same goes the other way: an average that held while the individual sessions swung is reported as no clear direction, not as holding steady.
When the later sessions average below the earlier ones, the log says that in a sentence. A tool that only ever reports good news is a tool whose good news means nothing.
If a focus switched from beats per minute to clean passes, only the sessions using the most recent measure feed the comparison, and the count left out is shown next to it. Averaging the two would produce a confident number about nothing.
There is no account, no upload, and no network call. Sessions are stored in this browser only. Export writes a plain JSON file you keep, importing it adds those sessions back, and the clear button removes the lot.
Pick whichever measure suits the thing you are working on. A passage with a speed ceiling wants a tempo — build the steps with the tempo ladder and log the rung you reached. Something that falls apart at random wants a pass rate: ten attempts, six clean, log 60. If you are unsure which failure you are looking at, the song readiness score names it, and practising with a metronome covers how to take a tempo reading you can trust.
What to run when the log says a focus has stopped moving:
Answer a short set of questions and name the part of your practice that stopped producing change.
Turn the minutes you actually have into timed blocks, weighted toward the one thing you are fixing.
Turn one difficult passage into a tempo ladder with a starting speed, a target, and steps between them.
Check a song against the parts that break under pressure before you call it finished.
What deliberate practice means at the instrument, and how it differs from playing something again.
Testing against the click instead of playing along with it, and what that exposes about your time.
How to run a 30-day challenge that ends with a recording you can compare, not a streak you kept.
The loop the whole site runs on: diagnose, prescribe, practice, prove, correct, repeat.
That part still takes another set of ears. The founding room is being assembled for it: a small crew, a weekly recording, and one correction that changes the next session. Applying starts a fit conversation. It does not charge you or create a commitment.