A lookup table that lists, for every word, the documents containing it, so a search engine can find matches instantly instead of reading everything. It works like the index at the back of a book, which tells you which pages mention a topic. Almost every keyword search system, from web search to searching your email, is built on one.
Technical view
Maps each term to a postings list of document identifiers, usually with term frequencies and positions to support ranking and phrase queries. Queries are answered by intersecting or merging postings lists, and scoring functions such as BM25 use the stored statistics. Compression of postings, skip pointers and sharding keep it fast at scale; Lucene and the engines built on it are the common implementations. It excels at exact terms, names and codes, which is why hybrid retrieval pairs it with dense vector search that handles paraphrase. Learned sparse methods reuse the same structure with model-assigned term weights.
Related
Part of AI Terminology Index, section: Retrieval and Knowledge.