Documentation
The notes that follow you home
Charting was supposed to save clinicians time; instead it became the job. Here is how we design software that gives the evening back.
The evening shift nobody agreed to
Ask any clinician what they did last night and a lot of them will tell you the same thing: notes. The clinic closed hours ago. The charting did not.
This is often called pajama time, and the name is doing a lot of polite work. It is unpaid work that follows people home, and it adds up.
The frustrating part is that documentation was meant to be the fix. Structured records, searchable histories, defensible audit trails. All good things. But somewhere along the way the record stopped being a byproduct of care and became a second job layered on top of it.
Where the friction actually lives
Most charting software makes you satisfy the data model before you can write a word. Want to record a phone call? Open a visit. Need to correct a note from Tuesday? Open a visit. Clinicians end up creating encounters that never happened, just to have somewhere legitimate to type.
That is a small indignity repeated forty times a week. It also pollutes the record: now the chart contains visits that were administrative fiction, and anyone reading it later has to work out which ones were real.
LogBlues lets you write a note straight onto a patient chart. No visit required, no check-in to fake. If something happened and it belongs in the chart, you write it in the chart.
Notes that do belong to a visit behave the way you would expect. Symptoms, the clinical note, and recommendations stay attached to that visit, and they stay editable afterwards. You are not locked out of your own thinking because a session was marked complete.
An assistant that drafts, not one that decides
The obvious pitch right now is AI that writes your notes. The obvious problem is that nobody sane wants a model silently committing text to a legal medical record.
Blues, our assistant, drafts. It can put together a task, a message, or a note, and then it opens the real form with the fields already filled in. Nothing is ever saved without you confirming it. You read it, you fix it, you commit it. The clinician stays the author.
Blues also runs on the clinic's own machine, on a local model runtime. Dictating a note or drafting a message does not ship PHI to somebody else's API endpoint. That is not a policy promise about how a vendor handles your data; it is an architectural one about whether the data leaves at all.
This matters more than it sounds. The moment a third-party API is in the loop, you inherit their retention, their subprocessors, and their outage schedule. We would rather not hand you that.
Local-first, because the record is yours
The same principle runs through the rest of the app. Notes are written and stored on the practice's own machines. Local-first is not a nostalgia play; it is what lets the software be fast, work when the connection does not, and keep PHI where you can actually point at it. Designed for HIPAA compliance from day one, not retrofitted after the first audit.
Follow-up work gets the same treatment. Tasks are one line each, showing who owes what by when. Completed ones stay visible and crossed off until they are swept away, so the day's work reads as a record rather than vanishing the instant it is done.
- Write where the thought happens. Chart-level notes without a manufactured visit.
- Keep editing. Visit notes stay attached and stay changeable.
- Confirm everything. Blues fills the form; you decide what is saved.
- Keep the data home. Local model runtime, local storage.
None of this eliminates documentation. Nothing will, and anything claiming otherwise is selling something. What it can do is stop the software from adding steps that only exist to serve the software, so the notes get finished while you are still at the clinic.