TinySearch v0.6.2

TinySearch v0.6.2

TinySearch v0.6.2 makes web research more focused and efficient, with smarter handling of overlapping crawl evidence and a set of quality and performance optimizations.

What's changed

• Semantic deduplication for ranked crawl chunks: TinySearch now removes semantically overlapping chunks from ranked crawl results, helping agents spend their context window on distinct, useful evidence instead of repeated passages. (#51)
• Quality and efficiency improvements: A broader optimization pass improves the search and research experience while keeping TinySearch focused on high-quality, source-grounded results. (#60)

A huge thank you to @benmaster82

This release is meaningfully better because of @benmaster82's contribution. The semantic deduplication work directly improves the quality of ranked crawl results and makes TinySearch more useful for every agent that relies on it for evidence-rich research. Thank you, Ben, for the thoughtful implementation and for helping push TinySearch forward. We really appreciate it.

Contributors

• @benmaster82 in #51
• @MarcellM01 in #60

Full Changelog: v0.6.1...v0.6.2