Healsens Platform

Health records, turned into structured, usable data

Healsens reads lab reports and health records, codes every result to a common standard and makes the data available to your software and AI agents. Confidence flags and human-in-the-loop review keep it reliable.

DOCUMENT WORKSPACE / LAB-REPORT.PDFIllustrative

Source document

Structured results

Haemoglobin14.2 g/dLLOINC 718-7 · high confidence

Glucose, fasting94 mg/dLLOINC 1558-6 · high confidence

Vitamin D (25-OH)Needs reviewValue unclear on scan · sent to reviewer

TSH1.8 mIU/LLOINC 3016-3 · high confidence

3 results ready · 1 waiting for reviewOpen review

Built on open standardsHL7® FHIR® R4 · R5LOINC®Model Context ProtocolGDPREU hosting for projects

What we do

One platform for the whole health-data path

From the first scanned page to answers and analytics. Each capability works on its own and runs on the same production platform as the Healsens app.

Extraction

Medical Document Processing

Lab reports and records as PDF, scan or photo, turned into structured results with a LOINC® code and unit for every value.

Learn more

Open source

MCP Server for FHIR®

Give AI agents safe, read-only access to an HL7® FHIR® server, with knowledge of what that server supports.

Learn more

AI assistant

Health Record Assistant

Answers questions from a person’s own record, built with retrieval-augmented generation.

Learn more

Analytics

Health Data Analytics

Explainable scores from published methods and trends over time, with every calculation visible.

Learn more

Services

Health Data Engineering

Platforms, integrations and AI features for health data, built and run by the team behind Healsens.

Learn more

Document intelligence

Process more documents. Type fewer results by hand.

Every lab lays out its report differently and names tests in its own way. The platform reads the page, understands the table and maps each value to a coded test, so results from different labs line up.

Medical Document Processing

OUTPUT / RESULTS.JSONIllustrative

{
  "document": "lab-report.pdf",
  "results": [
    {
      "test": "Haemoglobin",
      "loinc": "718-7",
      "value": 14.2, "unit": "g/dL",
      "confidence": "high",
      "review": "not needed"
    },
    {
      "test": "Vitamin D (25-OH)",
      "loinc": "62292-8",
      "value": null,
      "confidence": "low",
      "review": "pending"  // human-in-the-loop
    }
  ]
}

01Layout-independent extraction

OCR and a language model read tables and free text without a template per lab.

02LOINC® coding and unit normalisation

Each value is matched to a catalogue test and converted to a common unit.

03Confidence flags and human-in-the-loop review

Low-confidence values are flagged and routed to a reviewer before they are used.

04Delivery in the format you use

JSON through an API, CSV, or HL7® FHIR® Observation resources.

How it works

From a scanned page to data your systems can use

One pipeline for a single upload in an app and for a full archive of records.

STEP 01

Ingest

PDFs, scans and phone photos arrive through an API or in batches.

STEP 02

Recognise

OCR reads the page; a language model identifies tests, values and units.

STEP 03

Code

Values are matched to LOINC® codes and normalised to common units.

STEP 04

Review

Confidence flags send uncertain values to a person for review.

STEP 05

Deliver

Structured results flow to your API, files or FHIR® server.

Our AI

Practical AI, designed for health data

We combine proven components and keep a clear boundary around what each model may see and do.

Document understanding

OCR with Google Document AI and interpretation with large language models.

Human-in-the-loop

Confidence flags and review steps wherever an automated result could be wrong.

Agents and RAG

Assistants grounded in a person’s record and a curated knowledge base.

Open standards

HL7® FHIR®, LOINC® and the Model Context Protocol, so data and tools stay portable.

In production

The same platform runs the Healsens app every day

People upload their own lab reports, ask questions about their results and follow their health over time in the Healsens app for iOS and Android. Projects for organisations use the same components.

Industries

Built for teams that work with health data

Care and diagnostics

Clinics and laboratories

Structure incoming lab reports, digitise archives of paper and PDF records, and follow results over time.

Lab report intakeArchive digitisation

Research

Research consortia

Turn participants’ existing lab reports into coded study data and give analysts consistent access to it.

Study dataCohortsFHIR® access

Product

Digital-health companies

Add document upload, AI assistants and analytics to your product without building the foundations yourself.

APIsAI featuresApps

Trust

Privacy and security, by design

Access is limited to what each component needs. Agents read, they do not write. We agree the hosting and data flow with you before any data is shared.

GDPR

GDPRProcessing under EU data protection law

EU hostingProject deployments hosted in the EU

READ
ONLY

Read-only agentsAI agents read data, they do not write

Let’s talk about your data

Tell us what you are working on. We reply by email and, where it fits, suggest a short call to look at a sample together.