dailymed-query

name	dailymed-query
description	Query DailyMed for FDA drug label / package insert information. Use whenever the user asks about drug labeling, SPL documents, prescribing information, NDC codes, or needs to look up current FDA-approved drug details by name or NDC. Supports single entity or batch queries.

DailyMed Query Skill

Search DailyMed SPL labels by drug name or NDC code. No API key required.

Input Type	Function	Returns
drug name (str)	search_drug_labels(name, limit)	list of label summaries
SPL set ID	get_label_detail(set_id)	dict (parsed from XML)
NDC code	get_ndc_info(ndc, limit)	list of label summaries (JSON)
str or list[str]	query(entities, limit)	concise text summary
list[str]	search_batch(entities, limit)	dict[entity → results]

API

Function	Input	Returns
search_drug_labels(drug_name, limit=5)	drug name string	list[dict] with title, setid
get_label_detail(set_id)	SPL set ID	dict (XML→dict, keys vary by label)
get_ndc_info(ndc, limit=5)	NDC string	list[dict] with title, setid
search_batch(entities, limit=3)	list of drug names	dict[str, list[dict]]
summarize(results, entity)	results list + label	compact text
query(entities, limit=3)	single str or list[str]	LLM-readable text block

Usage

See if __name__ == "__main__" block in 06_DailyMed.py for runnable examples covering: single drug search, batch multi-drug search, label detail retrieval, and NDC lookup.

Quick Examples

from importlib.machinery import SourceFileLoader
dm = SourceFileLoader("dm", "06_DailyMed.py").load_module()

# Single entity
print(dm.query("metformin"))

# Multiple entities
print(dm.query(["aspirin", "lisinopril", "atorvastatin"]))

# Full label detail
hits = dm.search_drug_labels("metformin", limit=1)
detail = dm.get_label_detail(hits[0]["setid"])

Data Source

• Provider: U.S. National Library of Medicine / DailyMed
• Base URL: https://dailymed.nlm.nih.gov/dailymed/services/v2
• Auth: None (public API)
• Note: Search (/spls.json) and NDC lookup (/spls.json?ndc=...) return JSON. Detail endpoint (/spls/{setid}.xml) is XML-only; parsed to dict automatically.