Editorial policy
Who publishes MolForge, how its pages are produced, what they are (and are not) for, and how to get an error corrected.
Who publishes this site
MolForge (molforge.ai) is a molecular-discovery software platform. The public drug, protein-target and molecule pages are a free reference layer on top of the platform's database. Contact: hello@molforge.ai.
What the pages are for
The pages are for research and education: they summarise public scientific data about drugs, their protein targets and chemical structures. They are not medical advice and must not be used to diagnose, treat or make decisions about a patient. Always consult a qualified healthcare professional and the official prescribing information (for example on DailyMed or your national medicines agency).
Authorship and review
Data pages (/drug/, /target/, /molecule/) are compiled
automatically from the public databases listed in our methodology. Every
fact in their structured sections (indications, targets, approvals, adverse-event signals, identifiers,
properties) comes from a named upstream record, linked in the page's “Sources” section.
No page on MolForge is written or reviewed by a named clinician. We do not show author or reviewer names we cannot stand behind. If that changes, the reviewer will be named on this page and on the pages they reviewed.
Use of AI-generated text
- Drug guides (the long article on about 290 curated drug pages) were drafted with Google Gemini from the drug's name, structure, class and category. They are labelled as AI-drafted on the page. Their wording is automatically cleaned of filler; they have not been fact-checked line by line by a human. Where a guide and the DrugCentral data sections disagree, trust the data sections and the official label.
- AI insight panels on entity pages are generated by Google Gemini from the page's own structured data only, with instructions not to add facts that are not in that data and to say when the data is thin. They are cached and regenerated when the data changes.
- “Ask AI” answers are generated on request and are not reviewed.
- Frequently-asked-questions blocks marked as built from data are templated sentences filled with the page's structured data; no free-text AI is involved.
Dates and updates
Each data page shows “Data last refreshed”: the date its record was last loaded from the upstream source, read from our database. Guides also show when the article text was last edited. We never show today's date as an update date. Datasets are re-loaded when the upstream source publishes a new release; the per-source dates are on the methodology page.
Corrections
If you find an error, email hello@molforge.ai with the page URL and what is wrong (with a source if possible). We review every report. Errors in our own text (guides, summaries) are corrected or the text is removed; errors in upstream data are reported to the source database and, where we can, flagged on the page until the source is fixed.