DICOM-native preprocessing and routing
Every study normalised and sent to the model built for its modality, body part and protocol.
Latency · token cost · consistency
Applied medical AICDSCO Class B · Test-license
AI that fits into your radiology
and surgical workflows reliably.
Reporting intelligence that clinicians, patients and researchers can inspect, trust, adopt and improve.
Trusted by
The problem
Patient care quietly degrades when clinicians are stretched.
Radiologists
High volumes. Constant turnaround pressure impacts reporting quality.
Surgeons & endoscopists
Procedures back-to-back. Time for individualized patient care reduces.
Human errors
Fatigue, cognitive load, interruptions and delays compound.
AI errors
Unverified AI fails on incomplete and inconsistent real world data.
PixelMD Radiology Suite
Detect. Analyse. Report.
For radiologists. PixelMD supports in drafting standardized, actionable reports from imaging, dictation and patient clinical context across a wide range of modalities and pathologies.
See it in action
From dictation to a structured, checked draft — in real time.
app.pixelmd.ai/radiology · CT Brain · M. Kumar · 58M
LiveReading — CT Brain
00:01:00
Transcript
FindingsMeasurementsImpressionRecommendations
Receive
Detect
Analyse
Document
Quality control
Manual checks
Narrow pathology
Isolated reads
Free-text dictation
Manual proof-read
Auto-assign
Wide pathology
Reads with context
Actionable drafts
Automated QC
Receive
Manual checks
Auto-assign
Detect
Narrow pathology
Wide pathology
Analyse
Isolated reads
Reads with context
Document
Free-text dictation
Actionable drafts
Quality control
Manual proof-read
Automated QC
Diagnostic accuracy
Reading across the whole study to support accurate and consistent findings.
Turnaround time
Standardized drafting shortens the path from acquisition to a signed report.
Quality control
Automated checks on protocol, completeness and findings before sign-off.
PixelMD Surgery Suite
Analyse. Document. Assist.
For surgeons, PixelMD turns surgical or endoscopic video and voice into structured procedure documentation and post-operative monitoring plan. For duty medical officers and nurses, PixelMD assists with individualized in-patient management and discharge documentation.
See it in action
From ward-round voice notes to a complete draft — in real time.
app.pixelmd.ai/inpatient · Bed 12B · M. Kumar
LiveRecording — ward round
00:01:00
Transcript
VitalsMedsPlanInvestigations
Capture
Quality measurement
Documentation
Post-op monitoring
Handover & continuity
Limited access
No second pair of eyes
Op-notes from memory
Ad-hoc follow-up
Lost context across shifts
Fully queryable
Intra-op metrics
Op-notes from video
Timely intervention
Personalized records
Capture
Limited access
Fully queryable
Quality measurement
No second pair of eyes
Intra-op metrics
Documentation
Op-notes from memory
Op-notes from video
Post-op monitoring
Ad-hoc follow-up
Timely intervention
Handover & continuity
Lost context across shifts
Personalized records
Documentation quality
Structured op-notes and discharge summaries are claims and audit ready.
Post-operative monitoring
Personalized record across shifts and followups reduces medical errors.
Patient safety
Identifying complication patterns in early stage reduces readmissions.
The platform
Built by clinicians, for clinicians.
The same platform powers PixelMD Radiology Suite and PixelMD Surgery Suite — AI workflows, agentic quality control, deep clinical integration and AI output evaluation layer designed for radiology and surgical specialties.
01
Ingest
Image · video · voice · clinical context
02
Preprocess
Anonymize · normalize · augment
03
Model
Medical reasoning and reporting
04
QC checks
Protocol · findings · review
05
Output
Structured and traceable reports
Accelerators & FDE services
Embed the services, or have us deploy them.
Nine imaging services any PACS, VNA or EHR platform can call directly — and the team that puts them into production, across reads and documentation alike.
AI Accelerators · for PACS, VNA and EHR platforms
PHI redaction
Identity stripped from pixels, tags and frames, with recall measured per media type.
Compliance · governance
ROI annotation, segmentation and classification
Masks, measurements and labels returned as DICOM your viewer already renders.
Annotation effort · viewer capability
Worklist triage
The reading worklist re-ranked to each site's case mix, on your own scan-to-alert protocol.
Turnaround
Orchestrated analysis and drafting
Reads run on the smallest sufficient model; low-confidence sections escalate to a frontier one.
Compute cost · data residency
Priors and chart-context compression
Only the priors and chart sections that matter enter context — never the whole record.
Token cost · latency
Guardrails as an API
Any draft checked: sentences traced to source, laterality and dose errors stopped cold.
Report quality · governance
Prompt registry and eval loop
Each site's report format and language variant as versioned, testable configuration.
Onboarding speed · safe iteration
Opportunistic screening at archive economics
Existing archives re-read for incidental findings at a cost that makes screening viable.
GPU cost
FDE Services · for hospitals and groups
- 01ScopeWhich reads and documents, at what volume, in which systems.
- 02ConnectCredentials, callback, test round-trip per workflow.
- 03ShadowFull pipeline, nothing reaching a worklist or chart.
- 04Go liveExplicit, per site and per workflow. Your call.
- 05MeasurePixel-Proof, turnaround, acceptance, completeness.
Results come back in the standards your PACS and record already handle.
Pixel-Proof
A continuous evaluation layer behind every model run across every site we deploy.
Clinical Edit Distance
Critical Finding Capture
Polarity & Laterality
Hallucination Audit
Capture clinically important edits.
Capture what's dangerous to miss.
Right direction, right side.
Nothing invented. Nothing imagined.
Hedging & Confidence
Longitudinal Reasoning
Cross-Site Stability
Traceability
Uncertain at the right moments.
Reads the patient, not just imaging.
Every site, the same standard.
Full provenance for every output.
Clinical Edit Distance
Capture clinically important edits.
Critical Finding Capture
Capture what's dangerous to miss.
Polarity & Laterality
Right direction, right side.
Hallucination Audit
Nothing invented. Nothing imagined.
Hedging & Confidence
Uncertain at the right moments.
Longitudinal Reasoning
Reads the patient, not just imaging.
Cross-Site Stability
Every site, the same standard.
Traceability
Full provenance for every output.
On-premise
Runs inside your data centre. Data never leaves.
Private cloud
VPC deployment with regional data residency.
Hybrid
Imaging on-premise; reporting and audit in private cloud.
Standards & integrations
CDSCO Class B test licenseHIPAA / SOC 2DICOMHL7/FHIRRIS/PACSHIS / EHRSSO / SAMLAudit logs
The team
Clinicians and engineers, in the same room.
PixelMD began with a radiologist and a surgeon who lived the problem — thousands of reads and procedures under constant time pressure — and an AI engineering lead from pharma who knew how to build for regulated medicine. We build the tools we wished we had in the reading room and the OR.
Muthu Thangavelu
Chief Executive Officer
12+ yrs Pharma R&D and Commercial Ops, AI/ML Development. Nanyang Technological University, Merck, Daiichi-Sankyo.
Dr. Sathish Kandasamy, MD
Chief Radiologist
12+ yrs Image-guided procedures, Neuro, Vascular & Oncological Interventional Radiologist, Zulekha Hospital, Dubai.
Dr. Shyam Sundar, MS
Chief Surgeon
5000+ GI Laparoscopic & Robotic surgeries. CEO, Shaarve Hospitals; Sr. Consultant, RELA & Sugam Hospitals, Chennai.
The vision
Enhance diagnosis. Improve surgery. Return physician hours to patient care.
The best care should not depend on where you are. PixelMD exists to make this true.
By 2030 · every year
5 million
diagnoses enhanced.
500,000
surgical outcomes improved.
10 million
physician hours returned to patient care.
One standard of care. Every patient. Everywhere.
Contact
Partner with PixelMD.
Collaborate with us at the frontier of medical AI. Whether you are a hospital, tele-radiology or diagnostic network, biopharma or imaging CRO, tell us what you are building towards.