Plan: Fund · Credits: 1 per call; screen facets and the term search are free
The Product and Relationship Filters
string
Vocabulary search, not a screen — free. Returns product terms matching a spelling with how many companies make each. Call it before
makes_term: the vocabulary expands spellings, not concepts — memory, dram, nand and memory chips are four separate terms with four separate company lists, so makes_term=memory alone misses Kioxia, Winbond and Nanya. Pass every spelling it shows you.string
Comma-separated product terms; companies that make any of them. Exact terms from
find_terms.Companies related to any of the terms — in that product’s value chain without making it. Direction is usually unknown, so read it as “connected to”, not “supplies”.
string
Companies in this company’s disclosed compensation peer group, either direction. Takes a
company_id or ticker; an identifier matching nothing is an error, not an empty page.string
Companies named as a competitor by (or naming) this company in a 10-K Item 1. Same identifier rules as
peer_of.Every Match Carries Its Evidence
Matches come back withmatched_terms or matched_relation: the term, its sources, its grade, and whether a peer selection was mutual. The evidence travels with the row because extraction has failure modes — a business-section extraction can put “memory” on a PCB maker that merely wrote the word — and a product list nobody can check is a product list nobody should trust.