MedArchive
Turns a pile of clinic price lists in mixed formats into one searchable, normalized catalogue with an API.
- Period
- 2026
- Organisation
- MedPartners hackathon case
- source archives (ZIP, PDF, DOCX, XLS, scans)
- 10
- source archives (ZIP, PDF, DOCX, XLS, scans)
- price rows extracted
- 8,877
- price rows extracted
- services in the catalogue
- 6,614
- services in the catalogue
Clinics publish prices in whatever format they have: spreadsheets, Word files, scanned PDFs. MedArchive ingests them, matches each line to a service catalogue, and publishes the result through an API, exports and webhooks.
Stack
FastAPI and Celery for ingestion, a Next.js admin interface, OpenAPI/Swagger and Mintlify documentation.
Outcome
The submission reports 8,877 price rows and 6,614 catalogue entries from the organizers' 10-file archive. The live dashboard (27 Oct 2026) shows 18,901 extracted positions for 10 documents from a later processing run; the submission figures are the ones on record.
Evidence
01Built-in documentation of the pipeline