Your labs,finally alive.
Results arrive as scattered PDFs with no memory of the last draw. Labloom turns them into one bright record: what changed, what drifted, and what deserves a better question at your next visit.
Local-first, duplicate-safe, and built for real PDFs.
Latest read
Meridian 2025
Creatinine
1.28 mg/dL
HDL
51 mg/dL
Vitamin D
22 ng/mL
Reports
4
Markers
38
History
4y
Workflow
From scattered PDFs to one appointment-ready record.
The path is intentionally short: add reports, review what changed, and carry the clean summary into the visit.
Drop in the PDFs
Any lab, any year, any layout. Marker names, units, and date formats are reconciled — S-Creatinine in µmol/L and Creatinine in mg/dL become one line on one chart.
See what actually changed
Every value is checked against the range printed on its own report. What's outside, what moved more than 20%, and what's been drifting for years is stated up front.
Walk in prepared
Flagged markers become a printable list of questions for your next appointment — with the numbers and dates attached, so nothing gets hand-waved.
Privacy is the architecture
There is no server to send your bloodwork to.
PDFs are parsed in the browser and stored in this browser's IndexedDB. Labloom ships as static files — there is no backend, no account, and no analytics.
The one possible outbound request goes to Anthropic's API, using a key you paste yourself, and only for two optional jobs: reading a layout the built-in parser can't, and drafting questions from flagged markers. Neither runs without a key.
Export your record to a file whenever you want, and wipe everything with one control. Deleting really deletes — there is no copy anywhere else.
Labloom reports what your labs printed. It does not practise medicine.
Every flag is arithmetic against a reference range your own lab printed — never an interpretation, a diagnosis, or a recommendation. The output is a better-prepared conversation with your doctor, not a substitute for one.
Open Labloom