semantic_search_codespark 0.1.1

On-device semantic & vector search for Flutter — match by meaning, fully offline, no API keys. AI-powered search with a three-line API.

0.1.1

• Docs: rewrote the README around the use case (lead example, real-world object search, when-to-use guidance, limitations & tips).
• Discoverability: sharper description and topics (semantic / vector search, embeddings, AI, offline).

0.1.0

Initial release — offline semantic search for Flutter.

• SemanticSearch: one-call API to rank a list of strings or typed objects by meaning, fully on-device (no API keys, no cloud).
• search() / searchObjects<T>(): one-shot search with topK and an optional similarity threshold.
• createIndex() → SemanticIndex: embed a corpus once and query it repeatedly; supports diversity-aware ranking (MMR) via searchDiverse().
• SemanticResult<T>: carries the matched item, its cosine score, and original index.
• Pluggable SearchEmbedder (default CodesparkBackend over ai_core_codespark) so search logic is testable without downloading a model.