WonderSearch launches query-time document search without pre-embedding
The September 28 launch announcement says WonderSearch can search millions of unstructured documents without pre-embedding the corpus. It claims retrieval recall comparable to or better than embedding-based search.
TLDR
The September 28 launch announcement says WonderSearch searches at query time, with no corpus-wide embedding step or vector index to maintain. It claims comparable or better retrieval recall than embedding-based search and offers pay-per-search pricing instead of charging for embeddings that may go unused.
Combined views
1K
1 Source, first seen 7h ago
