Physicians & practices
Convert incoming laboratory reports into usable data and reduce repetitive capture across clinical and administrative teams.
Clinical Data Ingestion Engine
Clinical documents.
Structured data.
Powered by AI.
Turn laboratory reports and unstructured clinical documents into normalized, machine-readable healthcare data.
Less manual data entry. More time for care.
A foundation for connected clinical workflows.
Working laboratory ingestion MVP · Built with Gemini 2.5 Flash
LIVE MVP TESTER
GEMINI 2.5
Interactive front-end preview · Connected to live extraction engine
AI-powered extraction
LOINC mapping, where applicable
Original-source traceability
CSV / Excel outputs
01 / The problem
Clinical data should move.
Not be retyped.
Patients still bring printed laboratory results to their appointments. Critical clinical information remains trapped on paper, in PDFs and in scanned reports from multiple laboratories.
The time and effort physicians and their assistants spend on manual capture is invaluable and could be better devoted to patient care. Manual transcription is also inherently prone to human error.
01
Time taken from patient care
Repetitive capture and transcription add work for physicians and their teams.
02
Errors that travel downstream
Manual copying creates opportunities for missing or incorrectly entered information.
03
Information that stays fragmented
Different layouts and reporting conventions delay usable, structured clinical data.
02 / The solution
One pipeline.
From document to usable data.
CDIE uses AI to identify, extract and structure clinical information. Laboratory observations can be normalized and mapped to LOINC, where applicable, while retaining original values for traceability.
-
01 / INPUT
Clinical Document
PDF / Scan
Laboratory reports -
02 / EXTRACT
Gemini AI Extraction
Tests, results, units
Multimodal understanding -
03 / NORMALIZE
Normalize + LOINC
Standard terminology
Mapping where applicable -
04 / CHECK
Validate
Quality checks
Source traceability -
05 / OUTPUT
Structured Clinical Data
CSV / Excel · JSON
API in development
Pipeline overview. Expanded validation and confidence scoring are in development; interoperability integrations are on the roadmap.
03 / Working MVP
From a laboratory report
to structured observations.
The current MVP focuses on laboratory data ingestion: values, units, reference ranges, normalization and LOINC mapping, with original-source traceability.
01 / SOURCE DOCUMENT
PDF / Scan
02 / STRUCTURED OUTPUT
Excel + LOINC
Actual MVP screenshots supplied by the founders. These examples contain patient identifiers; replace with de-identified copies before public publication.
04 / AI at the core
Different documents.
A consistent representation.
AI is part of CDIE’s core architecture.
Laboratories use different layouts, terminology and reporting conventions. CDIE combines multimodal document understanding with normalization and validation to transform those heterogeneous inputs into structured clinical observations.
Current MVP
Google Gemini 2.5 Flash
Multimodal document understanding and clinical data extraction, combined with normalization and source traceability.
Planned evaluation
MedGemma on Google Cloud / Vertex AI
Evaluate healthcare-focused models for extraction, validation and clinical understanding as the platform evolves.
05 / Built for healthcare
One ingestion engine.
Multiple healthcare workflows.
Hospitals & networks
Build a foundation for standardized clinical information across document-based workflows.
EHRs & digital health
Planned API access will bring document ingestion and normalization into existing platforms and applications.
Business model
Clinical data ingestion
as a service.
CDIE is being built as a SaaS platform for individual physicians and healthcare organizations, with API access planned for platforms and enterprise customers.
Usage-based and monthly subscription options are under evaluation. Pricing and processing limits have not yet been finalized.
06 / Product roadmap
A working foundation.
A clear path forward.
Today’s laboratory ingestion MVP is the starting point for a broader clinical document intelligence platform.
Laboratory ingestion MVP
- AI-powered laboratory report extraction
- Structured observations, values, units and reference ranges
- Clinical normalization and LOINC mapping
- Original-source traceability
- CSV / Excel and structured data outputs
Production-ready engine
- Expanded validation and quality controls
- Confidence scoring
- Broader laboratory format coverage
- API-based ingestion and retrieval
- Production-grade security and operational controls
- Scalable cloud deployment
Healthcare interoperability
- FHIR-compatible structured outputs
- SMART on FHIR integration pathways
- HL7 v2 ORU/API integration pathways
- EHR / EMR connectors
- Additional clinical document types
- MedGemma evaluation on Google Cloud / Vertex AI
- Advanced clinical validation and enterprise ingestion workflows
07 / About CDIE
Healthcare interoperability
starts before the API.
CDIE — Clinical Data Ingestion Engine — is a healthcare technology startup building the infrastructure that turns human-readable clinical documents into machine-readable clinical data.
Based in Mexico City, Mexico, our mission is to eliminate unnecessary manual clinical data capture and make information usable by physicians, healthcare organizations and digital systems.
Luis Ángel Rodríguez Alemán
Co-Founder
Technology executive with CIO experience leading digital transformation across healthcare, banking, government and education. A computer engineering graduate of ITAM, Luis Ángel brings a cross-sector perspective on making technology serve real operational needs.
LinkedIn ↗
Jose Luis Jimenez Castillo
Co-Founder
Systems engineer with more than 10 years of software development experience and a background in Python solution architecture. Jose Luis brings software engineering and architecture expertise to CDIE.
LinkedIn ↗Let’s make clinical data usable
Upload a clinical document.
See the structured output.
Test the live extraction engine with a laboratory report in PDF or image format.
luisangelro@cdie.io · Mexico City, Mexico