Test phases
Sit-to-stand, walk out, turn, walk back, final turn and sit down, with the start, end and duration of each. The total time no longer depends on the reflexes of whoever holds the stopwatch.
In development, in pre-clinical validation
The Timed Up and Go, measured from a smartphone video.
Record the test with a phone and get, within minutes, the time of each phase, gait and balance metrics, and a clinical report ready to review and share.
We only send project milestones: validation results, test releases and collaboration opportunities. No spam, and you can leave at any time.
0.00s
Getting ready
Movement detected in a test recording, with no image of the patient
What it is for
The Timed Up and Go is the most widely used test of mobility and fall risk in older adults. It usually yields a single number: the total time. TUGscope turns the same 20-second recording into a complete gait and balance assessment, with no sensors, no markers and no lab.
Total and per-phase time
Sit-to-stand, walk out, turn, walk back, final turn and sit down, with the start and end of each phase. The stopwatch no longer depends on the examiner's reaction time.
Gait and balance metrics
Cadence, speed, step and stride length, stance and swing, double support, symmetry, regularity, turning strategy, trunk control and use of the hands during transfers.
A report ready to share
A clear PDF with a timeline, tables and a clinical narrative drafted with AI support. The clinician reviews, edits and signs before sharing.
How it works
All you need is a phone on a tripod. The video is analysed on your computer and never leaves it; only the measurements, and optionally a few key frames, are used to prepare the report.
Record the test
With a phone, from the side or the front, from the moment the patient is seated until they sit back down.
1 minuteLoad the video
Enter the patient, the walking aid and the course. The application tracks the body's movement and identifies each phase of the test.
On your computerGet the measurements
Total and per-phase time, cadence, speed, step length, turning, sit-to-stand and sit down, posture and signs of difficulty.
A few minutesReview and share
The report appears in the application with a clinical narrative proposed by the AI. Correct whatever you see fit and export the PDF.
PDF readyWhat it measures
TUGscope tracks the body's movement in every frame of the video and, from there, computes the measurements with reproducible biomechanical methods. Every value in the report states how it was obtained and with what confidence.
Gait
Step count, cadence, speed, step and stride length, stance and swing times, left–right symmetry and regularity.
Turn
Duration, number of steps, direction and smoothness of the turn, with warning signs such as pauses, crossed steps or en-bloc rotation of the body.
Sit-to-stand and sit down
Time of each transfer, number of attempts, use of the hands, trunk lean and control of the descent onto the chair.
Posture and signs of difficulty
Trunk lean and sway, arm swing, foot clearance, corrective steps and reaching for support on walls or furniture.
Recording quality
The application assesses framing, camera stability and detection confidence, and flags when a measurement is not reliable under those conditions.
First results
Total time measured by TUGscope against the examiner's stopwatch, in four test recordings under different conditions. The stopwatch includes the examiner's reaction time.
These figures come from internal development recordings. Formal validation, with a defined protocol and sample, is the project's next step.
| Recording conditions | Stopwatch | TUGscope |
|---|---|---|
| Handheld camera, following the patient | 27.13 s | 27.32 s |
| Frontal view, patient partly out of frame | 24.74 s | 23.6 s |
| Diagonal view, recording starts mid sit-to-stand | 16.11 s | 15.01 s |
| Frontal view, static camera | 13.26 s | 12.74 s |
The application
Three screens, no learning curve: patient details and video, analysis with the detected movement in real time, and results with a timeline, measurements and an editable report. Works offline; AI support is optional.
The PDF report
A document written to be read by other health professionals. It opens with the total time and the phase timeline, continues with the measurement tables and closes with the clinical narrative.
Measurements, observations and interpretation are always kept apart, so it is clear what was computed, what was seen and what is clinical judgement.
Missing values are declared as not assessed, with the reason: camera angle, no scale reference or too few steps.
A limitations section closes every report: video quality, camera stability, detection confidence.
Editable before export. The AI text is a proposal; the clinician signs what goes out.
See a real report (PDF, in Portuguese)
Report of the test shown in the video, exactly as exported from the application: 6 pages, 330 KB, in Portuguese. Published with the participant's consent; identifying details are anonymised.
Timed Up and Go test reportexemplo.mov · generated on 2026-10-28
12.7s
Total TUG time, from the start of movement until the hips settle on the chair
Sit-to-stand 0.5 s
Walk out 3.9 s
Turn 3.0 s
Walk back 2.6 s
Final turn 1.9 s
Sit down 1.1 s
PatientSample patient
Age78
Walking aidnone
Course3.0 m
Camerastatic
Camera viewdiagonal
Metric calibrationunavailable
Subject detection100% of frames
Gait
| Metric | Walk out | Walk back |
|---|---|---|
| Steps | 5 | 4 |
| Cadence | 109.7 steps/min | 131.5 steps/min |
| Speed | 0.77 m/s < 0.8 | 1.17 m/s |
| Step length | 0.42 m | 0.53 m |
| Regularity: step-time CV | 6.3 % | 35.9 % |
Turn
| Duration | 3.02 s | Pauses | 0 |
| Steps | 4 | En-bloc rotation | no |
Overall result. The total time of 12.74 s sits at the borderline of the cut-offs usually cited for higher fall risk in community-dwelling older adults (12 to 13.5 s) and does not, on its own, allow the risk to be classified. Walk-out speed (0.77 m/s) is below the 0.8 m/s reference threshold…
Continues: sit-to-stand and sit down, posture, signs of difficulty, interpretation, recommendations and limitations
With artificial intelligence support
The AI measures nothing: that is done by the movement-analysis algorithms. It comes in afterwards, to do what those algorithms do poorly: look at the images the way a clinician would, and write a clear report from the numbers.
Observe
Visual review of the test
From a few key frames of the video, the AI describes what can be seen: whether there is a walking aid, whether the patient reaches for support, whether the hands are used to stand up or sit down, whether the feet shuffle, what the posture looks like. When something is not visible, it says so.
Write
A clinical report in English
With the measurements and the visual review, the AI drafts a clear narrative, with cautious clinical references and a limitations section. It always separates what was measured, what was observed and what is interpretation. It makes no diagnoses.
Stays on your computerThe original video, the patient's identity, the video with the detected movement and the final PDF.
Shared for the reportThe measurements and, if you choose, a few key frames of the test. No patient name and no full video. Sending frames can be switched off.
Comes back to youThe observations and the report text, ready to review in the application before export.
Numbers are always measuredThe AI never invents or computes values. Every number in the report comes from the movement analysis and can be traced back to its source.
OptionalThe full analysis works without AI. The assisted report is switched on when you want it, and can be switched off at any time.
The clinician has the last wordThe text is a proposal. It is reviewed and edited in the application by whoever signs the report.
Project status
Software in development, in pre-validation for clinical studies.
TUGscope is not yet a certified medical device and must not be used for clinical decisions. The prototype works end to end and is ready to be compared with the reference assessment in clinical and hospital settings.
- Agreement and test–retest reliability studies in geriatrics, physical and rehabilitation medicine, neurology and orthopaedics.
- Standardised recording protocols (tripod, side view, 3 m mark) and collection of videos with informed consent.
- Collaboration with physiotherapists and geriatricians to review the clinical report and the references used in its interpretation.
Clinical and academic partners
If you work in a hospital, a long-term care unit, a teaching institution or a research group and would like to take part in the validation, get in touch.
[email protected]Diogo Pinto, Geridoc. We reply to every message.