0.1.7

Diff: 0.1.6...0.1.7

Added

• Related Notes View now uses the editor's text selection as the search query
  when in auto-follow mode. Works in both editing and reading modes; clearing
  the selection returns to the existing context-following behavior. A small
  "from selection" badge is shown next to the query header while a selection
  drives the query. For selection-driven queries the token budget is widened
  from maxQueryTokens to the embedder's actual context window (n_ctx
  reported by llama-server's /props endpoint), so an explicit selection can
  use the full model context.
• Related Notes View toggle button (and matching Include title in query
  setting) for prepending the active note's title to the search query. The title
  is included by default, matching previous behavior.

Changed

• Frontmatter is now always stripped from the active note before building the
  related notes search query, so YAML fields no longer leak into the query when
  the cursor or viewport is near the top of a note. To include frontmatter
  content in the query, select it explicitly.

• Default chat model updated from Qwen3-8B to Qwen3.5-9B
  (unsloth/Qwen3.5-9B-GGUF). Users may need to upgrade llama.cpp to the latest
  version for Qwen3.5 support.