HealthHub Screenshot Extraction

Pull Lab Results, Vaccine Records, and Appointments from a HealthHub Screenshot

Singapore's national health app lets you view and download lab reports and immunisation records, but each download is a document, not a spreadsheet. Custom Column Extraction reads the fields you name from a screenshot.

5-10s per screenshot · Up to 99% accuracy on clear screenshots

Test Name · Result · Range
Vaccine Name · Date
Appointments

What You Can Pull from a HealthHub Screenshot

HealthHub pulls lab results, vaccine records, and appointments from Singapore's National Electronic Health Record (NEHR) system. You can open and download a record, but nothing lines up in one table. With Custom Column Extraction, you name the fields and the AI reads them off the screenshot.

Test Name

The biomarker or test name, such as "Glucose, Serum" or "HbA1c", recorded as it is printed on the report.

Result Value

The measurement with its unit, such as "5.7 mmol/L" or "102 mg/dL", read even when it sits between the test name and the reference range.

Reference Range

The range the laboratory printed beside the result, such as "3.9-5.5 mmol/L" or "4.0-5.6%". Kept in its own column so results from different visits stay comparable.

Vaccine Name and Date

The vaccine, the date it was administered, and the batch number, such as "HPV (Gardasil 9), 12 Jan 2025, Lot #W1234", each extracted as its own field.

Appointment Details

Facility, department, date, and time, such as "SGH, Cardiology, 15 Mar 2025, 10:30 AM". HealthHub consolidates appointments across the public healthcare clusters.

Singapore's National Health Platform: Records Are There, but Not in Rows

Why HealthHub Records Are Hard to Put in a Spreadsheet

01

Each record comes out as a document, not rows. You can view and download a lab test report or an immunisation record, but what you get is a document built for reading. There is no way to pull selected fields from several records into one sortable table, or to merge your own records with your child's.

02

Records are spread across cluster apps. Singapore's public healthcare runs three clusters: SingHealth (Health Buddy), NUHS, and NHG Health. HealthHub is the national aggregator, but each cluster kept its own app with a different layout and login. They are being folded into one enhanced HealthHub app by November 2026.

03

Screens from different record types do not line up. A lab test report shows different fields from an immunisation record. A parent managing their own and their child's records has to capture several screens across several profiles and assemble them by hand.

How Semantic Extraction Reads Any HealthHub Screen

01

Name the fields and the AI finds them by meaning, not pixel position. Define "Test Name, Result Value, Reference Range." Whether a lab report places the range to the right of the result or below it, the AI locates each value by its role, the same way you can tell "5.7 mmol/L" is a result and "3.9-5.5 mmol/L" is the range.

02

Mix record types in one batch. Upload lab test reports and immunisation records together. The same column structure handles both: Test Name, Result, and Range for lab screens, Vaccine Name, Dose, and Date for immunisation screens. The output is one spreadsheet.

03

No template setup, no training data. Unlike enterprise OCR tools that need you to draw bounding boxes for each screen layout, Custom Column Extraction works with zero configuration. Upload a screenshot, type your column names, and the AI adapts to whatever layout HealthHub or a cluster app displays.

From HealthHub Screenshots to a Consolidated Medical Record

1

Screenshot HealthHub Screens

After you sign in with Singpass, open the lab test report and take a screenshot. Open immunisation records and take another. If you manage a child's profile through caregiver access, screenshot theirs too. Each capture is a clean image, and HealthHub's interface is well-rendered, so the text holds up well for extraction.

2

Name Your Columns Once

Type the field names you need: Test Name, Result Value, Reference Range, Vaccine Name, Date Administered, Appointment. Upload all screenshots into a single batch. The AI reads each image independently: from a lab report it captures values such as glucose and HbA1c; from an immunisation record it reads the vaccine name and date, all under the same column structure.

3

Download a Unified Spreadsheet

Processing takes about 5-10 seconds per screenshot. The output is a single Excel (XLSX) file: rows from your lab results, your immunisation records, and your child's vaccination history, all aligned under the headers you defined. That is roughly 18x faster than copying each field from the screen by hand, which takes about 3 minutes per screenshot.

When It Works Best, and When to Be Cautious

HealthHub's clean interface suits screenshot extraction, and a few boundaries are worth knowing.

When It Works Best

✓

Full-resolution screenshots from the HealthHub app or web portal. Machine-rendered medical data with high contrast reaches up to 99% accuracy on text fields.

✓

Records from any NEHR-linked app. HealthHub, Health Buddy, and the web portal all draw on the same national records. Screenshots from any of them work with the same column definitions.

✓

Caregiver multi-profile batches. Screenshots from your profile and your child's profile upload together into one consolidated family spreadsheet.

When to Be Cautious

⚠

Coverage depends on where the test was done. HealthHub retrieves lab and screening records from participating public healthcare institutions such as polyclinics and restructured hospitals. A result from a private clinic or an overseas lab may not appear there at all, so check the source app before you extract.

⚠

Long lists spread across screenshots. A lab list or appointment history that runs past one screen needs several captures. Each screenshot processes independently under the same column headers.

⚠

Visible data only, with no clinical interpretation. The tool extracts what the screen shows, not what it means. Share the spreadsheet with your GP for clinical decisions.

Frequently Asked Questions

Can I extract both my own and my child's HealthHub records in the same batch?

Yes. HealthHub supports caregiver profiles, so you can view your child's immunisation records and lab results from the same app. Screenshot each profile and upload them together. Define columns like Test Name, Result Value, Reference Range, and Vaccine Name once. The AI processes each screenshot independently and outputs one merged spreadsheet with every family member's data.

Does it capture the reference range the way HealthHub shows it?

Yes. A lab test report lists each test with its result, unit, and the reference range the laboratory printed beside it. The Reference Range column captures that exact text, such as "3.9-5.5 mmol/L" for fasting glucose or "4.0-5.6%" for HbA1c, and keeps it separate from the result value so you can compare visits against the same standard.

Can I extract vaccine batch numbers for overseas travel documentation?

Yes. HealthHub's immunisation records show the vaccine name, date administered, batch number, and administering facility. Define those as your columns and the AI reads them from the screenshot. This helps with university enrolment or travel that requires batch-level proof, such as COVID-19 boosters or HPV vaccines listed in Singapore's National Immunisation Schedule.

What about screenshots from Health Buddy or other cluster apps?

They work too. Health Buddy (SingHealth), OneNUHS, and NHG Health all draw on the same NEHR records, so the field names stay consistent even though the layouts differ. Singapore is folding these cluster apps into one enhanced HealthHub app by November 2026, and screenshots from any of them process with the same column definitions.

How accurate is extraction for HealthHub lab result screenshots?

For clear screenshots of HealthHub's on-screen lab reports, accuracy on numeric values reaches up to 99%. The AI reads the contrast-rich, well-spaced text those screens use. Heavily compressed images, such as a photo of a screen, reduce accuracy on smaller reference range values.

📮 contact email: [email protected]