He Yaoshi Drug Label JSON Toolkit

Problem

Biomedical data is fragmented across databases: ChEMBL covers drug bioactivity, OpenTargets covers target evidence, PubMed covers literature, ClinicalTrials.gov covers trials, and OpenFDA covers safety. For drug repurposing or target discovery, querying these sources manually means repeatedly handling terminology, ID formats, and filters, making it hard to build a reviewable chain from candidate drugs to evidence and trial status.

How It Works

The skill aggregates multiple public databases into a single query entry point for disease-, gene-, and drug-centric research:
- Find drugs: query approved and investigational drugs for a disease, with the maximum phase reached.
- Find targets: rank candidate targets by disease association and evidence strength.
- Find evidence: search literature, actively recruiting trials, and OpenFDA adverse-event data.
- Compose workflows: target discovery, drug matching, trial validation, and safety checks can be chained into reusable research queries.
Most endpoints are free, support batch queries and cross-database ID conversion; the production endpoint requires no authentication and uses automatic caching.

Boundaries

It supports literature review, candidate screening, and evidence synthesis, but does not replace clinical trial design, medical judgment, or regulatory review. OMIM requires an API key; some databases use different ID systems, so empty results may require standardized terms or filters such as max_results, trial phase, or recruitment status.