This design is optimized to reduce infringement and terms-of-service risk while still creating a useful disease-centric literature intelligence product.
1. Collect only what you can lawfully collect
Use API / RSS / allowed bulk services for PubMed, Europe PMC, bioRxiv, medRxiv, Crossref, and OpenAlex.
2. Store minimal content
Keep IDs, titles, author metadata, dates, badges, citations, journal metrics, and our own digested headlines.
3. Link out for reading
Send users back to the source record or publisher page instead of hosting the article PDF or publisher layout.
Store metadata, links, IDs, badges, and original in-house digests only.
Do not mirror PDFs, figures, or full abstracts unless the item license explicitly allows it.
Restrict automated full-text retrieval to open-license repositories and approved APIs.
Keep a takedown / contact workflow, source audit log, and per-item rights flag.