Super Adds Search Across Agent Chats

This is less “AI memory” than fast retrieval, but it solves a painful failure mode in agentic development: repeatedly rediscovering decisions already buried in chat history.

• –Full-text indexing turns past agent sessions into reusable engineering context
• –Archived-chat search is especially valuable for long-running repositories and recurring bugs
• –Command-palette shortcuts keep retrieval inside the developer workflow
• –Local-first history handling is a stronger privacy story than sending transcripts to a separate knowledge service
• –Search quality will ultimately depend on semantic retrieval, filtering, and conversation organization beyond keyword matching