No safety flag and β€10% of facts missed or wrong
API Testing Dashboard
Did not meet the pass rule
No safety concerns in this batch
Test Runs
Configure doctors and patients, then run tests.
1. Doctor & Patient Setup
Each doctor is linked to one patient for a test run.
| Username / Phone Number | Password | Patient ID | Action |
|---|
2. Upload Multiple Audio Files & Ground Truths
Upload multiple audio files along with their ground truth (JSON / Excel / PDF).
Drag & drop audio files and ground truth files here
or Browse Files
Supported: mp3, wav, m4a, json, xls, xlsx, csv, pdf
| # | AUDIO FILE | DURATION | GROUND TRUTH | STATUS | ACTION |
|---|
No audio files selected for this run. Run Batch Test is disabled until at least one file is listed.
3. AI Model & Settings
This will run all selected files and compare with their ground truth.
| Test Case ID | Audio File | Language | SOAP accuracy | Clinical Quality | Execution Status |
|---|---|---|---|---|---|
| No previous test cases yet | |||||
SOAP is Flask llm-time (SOAP/summary generation).
Missing timings show as unavailable.
Summary
Load Testing
Test system performance under load with multiple doctorβpatient sessions.
1. Doctor & Patient Setup β
Each row represents one doctorβpatient pair (1 to 1).
| Username / Phone Number | Password | Patient ID | Action |
|---|
Total Rows: 0
2. Audio & Ground Truth
Provide audio and ground truth files per patient.
e.g. 101_audio.wav / 101_ground_truth.txt
If a patient ID has no matching file, a silent WAV fallback will be used for audio and ground truth will be skipped.
3. Config (Optional)
Import or export your test configuration.