Lore 0.2.0 adds visible reasoning, non-destructive embedding migrations

Lore's 0.2.0 update highlights the maturing ecosystem of local-first AI tools where data privacy is the default, not an afterthought.

• –Exposing the agent's "ThinkingStream" builds developer trust by showing exactly how local notes are retrieved and processed
• –Non-destructive embedding migration solves a major pain point for local RAG systems, allowing users to upgrade embedding models without losing data or re-indexing from scratch
• –By pairing Ollama with LanceDB, Lore proves you can build a highly capable personal knowledge assistant without external API dependencies