Julie — the AI radiologist with the equivalent of 40 years of experience — on call 24/7.
Upload a CT, MRI, X-ray, mammography or ultrasound study. In under 60 seconds, Julie returns a structured, ACR/RADS-compliant report with findings, differential diagnoses, RADS scoring and recommended next steps — ready for your sign-off.
Meet Julie — your AI radiologist, in action.
Watch Julie read a thoracic CT, flag findings, and draft a structured report — in seconds. This is a live-style simulation of what she does inside the platform.
A complete reading workflow, not just an image classifier.
Every report follows the same discipline a senior radiologist would apply: technique, comparison, structured findings, impression and recommendations — backed by the most relevant medical standards.
Multi-modality reading
CT, MRI, X-ray, mammography, ultrasound, PET-CT. DICOM and standard image formats.
Structured ACR/RADS reports
BI-RADS, LI-RADS, PI-RADS, Lung-RADS scoring applied where clinically relevant.
Prior-exam comparison
Upload a prior study and the AI flags new, resolved or progressing findings.
Literature & guideline search
Differential diagnoses anchored in PubMed evidence and ACR Appropriateness Criteria.
Voice dictation
Dictate edits to the impression and recommendations in your own voice.
Translation in 12 languages
Translate full reports for international patients and multi-site teams.
Patient Mode
One-click plain-language summary written for patients, not clinicians.
Branded PDF export
Clean, hospital-grade PDF reports with your logo and clinician signature.
Always under your sign-off
Every report ends with a clear disclaimer — the final decision is yours.
Trained across the conditions clinicians actually read every day.
Julie is tuned on millions of public and de-identified imaging studies and reasoning traces, then aligned with veteran radiologist reading protocols.
- Chest: nodules, masses, consolidation, pneumothorax, effusion, embolism
- Neuro: stroke, hemorrhage, mass effect, MS lesions, atrophy patterns
- Abdomen: liver lesions (LI-RADS), renal/adrenal masses, bowel obstruction
- Musculoskeletal: fractures, joint effusions, soft-tissue masses
- Breast: BI-RADS lesion characterization on mammography & ultrasound
- Genitourinary: PI-RADS prostate assessment on multi-parametric MRI
From upload to signed report in under 2 minutes.
- 01
Upload study
Drag a DICOM, JPG or PNG into the workspace. Optional prior exam for comparison.
- 02
AI reading
Veteran-style protocol: technique → findings → impression → recommendations.
- 03
Refine & dictate
Edit any section, dictate by voice, translate, generate patient summary.
- 04
Sign & export
Branded PDF with your sign-off — ready for the EHR or to send to the patient.
Every task a senior radiologist performs — automated, structured, auditable.
Julie doesn't just "describe an image". She runs a complete reading protocol on every study, in the exact order a 40-year veteran attending would.
1 · Reading protocol on every study
- 01Identify modality, region, sequences/phases and technical adequacy
- 02Note patient context (age, sex, clinical question) when provided
- 03Compare against prior exam if uploaded — flag new / resolved / progressing findings
- 04Systematic anatomical review (e.g. chest CT: lungs → pleura → mediastinum → vessels → bones → upper abdomen)
- 05Measure lesions in mm, describe density / signal / enhancement, location with anatomical landmarks
- 06Apply the relevant RADS lexicon (BI-RADS, LI-RADS, PI-RADS, Lung-RADS, TI-RADS, O-RADS)
- 07Generate impression: ranked differential diagnoses with confidence
- 08Recommend next steps anchored in ACR Appropriateness Criteria
- 09End every report with the human sign-off disclaimer
2 · Structured report sections produced
3 · What she reads, by modality
- Pneumonia, consolidation, edema
- Pneumothorax & effusion
- Nodules / masses
- Cardiomegaly
- Line & tube placement
- Pulmonary embolism
- Lung-RADS nodule scoring
- Interstitial lung disease patterns
- Mediastinal & hilar nodes
- Aortic pathology
- Acute ischemic stroke (ASPECTS)
- Hemorrhage classification
- Mass effect & midline shift
- MS lesion burden
- Atrophy patterns
- LI-RADS liver lesions
- Renal & adrenal masses
- Bowel obstruction / ischemia
- Pancreatitis & ductal pathology
- Appendicitis & diverticulitis
- Fracture detection & classification
- Joint effusion & arthropathy
- Soft-tissue masses
- Spine compression & alignment
- Bone lesions (lytic / sclerotic)
- BI-RADS lesion descriptors
- Mass vs. calcification analysis
- Asymmetry & architectural distortion
- Density category (a–d)
- Final BI-RADS 0–6 assessment
- PI-RADS v2.1 scoring
- Peripheral vs. transition zone
- DWI / ADC correlation
- DCE assessment
- Extracapsular extension
- TI-RADS nodule scoring
- Composition, echogenicity, margins
- Calcification pattern
- Lymph node assessment
- FNA recommendation
- O-RADS adnexal scoring
- Endometrial assessment
- Fibroid mapping
- Endometriosis signs
- Cervical evaluation
4 · Tasks you can run on any report
- Compare with a prior study and auto-generate a delta report
- Pull supporting PubMed evidence and ACR criteria inline
- Dictate edits by voice into any section
- Translate the full report into 12 languages
- Generate a plain-language Patient Mode summary
- Export a branded hospital-grade PDF with your signature
- Audit trail of every AI suggestion vs. your edits
5 · What she will never do
- Issue a final diagnosis without a human radiologist's sign-off
- Bypass HIPAA / GDPR data handling — every study is encrypted and access-logged
- Use your patient data to train base models
- Replace a licensed physician — it is explicitly a clinical copilot
- Hide its confidence — every finding ships with a likelihood and reasoning
Built with radiologists, never to replace them.
“It reads like a senior attending writing the report — structured, cautious, and clinically anchored.”
“We cut average reporting time from 14 to 6 minutes per study. The ROI was obvious within a month.”
“Patient Mode alone changed how we communicate results. Patients finally understand their imaging.”