Keep a physical therapy log without typing. Capture sets, reps, pain scores, and notes by voice so your PT sees real progress between visits.
Your physical therapist gives you six exercises, a rep count, and a warning about which movements should hurt and which should not. By day three you cannot remember whether the band exercise was three sets of twelve or two of fifteen. By your next appointment two weeks later, you have forgotten which day your knee flared up. So you shrug and say "it's fine, I think," and your PT adjusts the plan on a guess.
A physical therapy log fixes that gap. It is a running record of what you did, how many reps, how it felt, and what changed. Most people abandon it within a week, because writing it down after a workout is one step too many. This guide covers what to record, how to keep it simple enough to stick to, and how to capture entries by speaking instead of typing.
Rehab runs on progressive load. You do a little more each week, and that only works if someone knows exactly where you were last week. Memory is a poor instrument for it: you will overestimate the good days and forget the bad ones.
A log turns your recovery into data your therapist can read. Instead of "it felt sore," you can say "left knee was a 4 out of 10 during step-downs on Tuesday, back to a 2 by Thursday." That one sentence tells a clinician whether to hold, progress, or regress the exercise. It also protects you from doing too much: a pain score climbing three days running is a reason to ease off before a small flare becomes a setback. Nobody claims a notebook heals tissue, but a clearer picture may support better decisions.
Keep the fields few. A log with fifteen columns gets abandoned; a log with five gets filled in. Here is a minimal PT log template that still gives your therapist enough to work with.
| Field | What to capture | Example |
|---|---|---|
| Date and time | When you did the session | Tue 22 Jul, 8:10am |
| Exercise | Name your PT used | Terminal knee extension |
| Sets and reps | What you actually did, not the plan | 3 sets of 12 |
| Pain score (0 to 10) | During and after the movement | 4 during, 2 after |
| Note | One line on form, fatigue, or a flare | Right side weaker, cut last set short |
Five fields. That is the whole thing. Two optional extras worth adding: resistance (band colour, weight, or machine setting) and sleep or stress that day, since both affect how a session feels. The best rehab progress tracker is the one you still open in week six.
The pain score column matters most. Rehab pain is not always bad pain, and your therapist needs the number, not the adjective. A steady 2 to 3 that fades after the session is usually fine; a 6 that lingers into the next day is information your PT wants immediately.
Ask anyone who quit their log why, and the answer is never "data was useless." It is "I stopped doing it." The session ends, you are tired, your phone is across the room, and typing five fields feels like homework. So you skip one day, then three, and the log is a graveyard of two entries.
The fix is to cut the time per entry from two minutes to twenty seconds. That is what speaking does. Right after your last rep, while the detail is fresh, you say the entry out loud and it lands in your file as text. No hunting for the app, no thumb-typing "terminal knee extension" three times a week.
The best physical therapy log is not the most detailed one. It is the one you still fill in during week six, when the novelty is gone and the friction is all that is left.
This is where a context-aware speech-to-text tool differs from generic dictation. Contextli is a desktop app that types into whatever window you have open, so your log lives in the note file or template you already use. No new inbox to check, no export step: you speak, and the text appears where your cursor is.
The workflow is short. Open the note or template where your log lives, put your cursor at the next empty row, and speak the entry. With Notes Mode, Contextli formats spoken rehab notes as clean, readable lines instead of a run-on transcript, so "three sets of twelve, pain four during, two after, right side felt weaker" comes out as a tidy record rather than a wall of words.
The part that saves you over time is customization. You can feed a Mode three or four examples of how you want entries to read, or a plain instruction like "always log pain as a number out of 10 and put the exercise name first." After that, every dictation matches that structure without you repeating the format. Your log stays consistent from the first entry to the hundredth, which is exactly what your therapist needs to read a trend.
Because health notes can be sensitive, the privacy setup is worth naming. Contextli has a stackable privacy ladder. First, local models: transcription and text processing run on your own machine, and you can turn the internet off with the app still working, so audio and text never leave your laptop. Second, bring your own key (BYOK): you supply your own provider API key and your data goes from your machine to that provider directly, with Contextli never seeing it. Third, disable cloud sync: turn it off and Contextli stores nothing in its database, so your notes stay as local files. Stack all three and your rehab log is fully offline and fully private. The honest trade-off with local models is they need a reasonably modern Mac or Windows laptop; an older machine runs them slowly.
Here is how the options stack up on the thing that decides whether you keep going: friction per entry.
| Method | Time per entry | Ends up in your own file | Notes stay private on device |
|---|---|---|---|
| Handwritten notebook | 60 to 90 seconds | Yes, on paper | Yes |
| Thumb-typing in a notes app | 90 to 120 seconds | Yes | Depends on the app |
| Otter.ai recording | Fast to speak, then export | No, lives in Otter first | No, processed in the cloud |
| Contextli voice, Notes Mode | About 20 seconds | Yes, types into your file | Yes, with local models on |
Otter.ai is a capable meeting transcription tool and records a full session well, but for a quick daily log it adds a step: your words live inside Otter and you have to go and get them. Contextli types the entry straight into the file you already keep, and with local models on, the recording and text stay on your machine. If all you need is raw transcription with no formatting or privacy concern, a simpler dictation tool may be enough. The reason to reach for a context-aware app here is the formatting plus the on-device option for health data.
Consider a runner six weeks into recovery from a knee strain. She finishes her home program in the kitchen and has a simple template pinned in a note file.
She puts her cursor on the next row and speaks: "Tuesday, terminal knee extension, three sets of twelve, pain four during and two after, step-downs felt shaky on the right, cut the last set to eight." Using Notes Mode, Contextli types it in as three clean lines. Total time from last rep to logged entry: about 25 seconds. The same detail thumb-typed on her phone would have taken close to two minutes, which is exactly the gap where she used to give up.
Two weeks later she opens the file at her appointment. Her therapist sees the right-side pain climbing on step-downs and swaps that exercise for an easier variation on the spot. No guessing, no "it's fine, I think." That is what a physical therapy journal is for.
At minimum: the date, the exercise name your PT used, the sets and reps you actually did, a pain score from 0 to 10 during and after, and one short note about form or how it felt. Five fields is enough for most rehab plans, and keeping it short is what makes the log survive past the first week.
Every session you do, including home programs, not just clinic visits. The point is to capture the days between appointments, since those are the days your therapist cannot see. Even a one-line entry per session gives a useful trend over two weeks.
Yes. A context-aware desktop app like Contextli types your spoken entry directly into your note file or template. Speaking an entry takes about 20 seconds versus one to two minutes of thumb-typing, which is the difference that keeps most people logging past the first week.
It depends on the app. With Contextli you can run local models on your own machine, turn the internet off, and disable cloud sync, so the recording and text never leave your laptop and nothing is stored in a server database. Always confirm any tool's data policy before logging sensitive health information.
Most therapists work with a 0 to 10 scale, where 0 is no pain and 10 is the worst you can imagine. Record the number during the movement and again after, since pain that fades quickly reads very differently from one that lingers into the next day. Use the number, not adjectives.
No. A plain note file with five columns works. The value is in filling it in consistently, not in the tool. If typing is the barrier, voice capture into that same file removes the step that usually makes people quit.
If you bring it to appointments, most will. A PT who can see two weeks of real sets, reps, and pain scores may adjust your plan faster and more accurately than one relying on your memory. Hand it over at the start of the visit.
These guides use the same voice-capture approach:
Contextli types your spoken PT entries straight into the note file or template you already keep, and with local models turned on, the recording and text stay on your own machine. Try it on your next session and see how a 20-second entry compares to two minutes of thumb-typing. You get 100 credits/month on the free tier, no credit card required. Download Contextli and log your first exercise today.
Junaid Khalid
Founder & CEO
Founder and solopreneur writing about how modern businesses run leaner and faster with AI. I build software that turns everyday work, from capturing thoughts to writing and staying organized, into something effortless, and I share what I learn along the way.
A recovery journal with daily prompts for people in recovery. What to write, how to keep it private, and how to capture entries by voice in under a minute.
Junaid Khalid
Build a recipe experiment log that captures what worked, what failed, and what to try next. Voice-first notes so you write while the pan is still hot.
Junaid KhalidCapture real estate showing notes by voice the moment you leave each property, so five homes in one afternoon never blur into one.
Junaid Khalid
