TextAid v1.3.0

Added

• Inline suggestions (ghost completions): when the toggle is enabled in the
  popup, TextAid listens for typing in any editable field (input, textarea,
  contenteditable). After 20+ characters and a 600 ms pause at a word
  boundary, it fetches a short AI completion and displays a dismissible widget
  below the field. Press Tab to accept and insert the suggestion, or
  Escape to dismiss it. Turning the toggle off immediately hides any
  pending widget.
• Three new keyboard shortcuts for direct actions on the current selection:
  Alt+Shift+S (Summarize), Alt+Shift+R (Rewrite), Alt+Shift+T
  (Translate to English).
• Free / Paid badge next to every model in the popup dropdown so it's
  obvious which Gemini models stay free and which require billing.
• Refreshed model lineup (May 2026):
  • Gemini: gemini-2.5-flash (new default), gemini-2.5-flash-lite,
    gemini-3-flash, gemini-2.5-pro, gemini-3-pro (legacy 2.0-flash
    and 1.5-pro kept for back-compat).
  • OpenAI: added gpt-4.1-nano, gpt-4.1-mini, gpt-4.1, gpt-5-mini,
    gpt-5 alongside the existing gpt-4o-mini / gpt-4o.
• Markdown rendering in the result modal: bold, italic, lists, code
  blocks, inline code, headings, blockquotes, links and --- rules now
  display as formatted HTML instead of raw markdown characters. Copy /
  Replace / Insert still use the raw markdown text.
• Translate to… submenu with 16 languages (English, Français, Español,
  Deutsch, Italiano, Português, Nederlands, Polski, Русский, Українська,
  Türkçe, العربية, 中文, 日本語, 한국어, हिन्दी) — replaces the old
  hard-coded "Translate to English" entry.
• Richer API error messages: parses Gemini (UNAVAILABLE,
  RESOURCE_EXHAUSTED, INVALID_ARGUMENT, PERMISSION_DENIED, …) and
  OpenAI (rate_limit_exceeded, insufficient_quota, model_not_found,
  context_length_exceeded, …) error bodies and surfaces actionable
  hints (e.g. "switch to a Flash / mini model", "add billing", "shorten
  the text"). Honors the Retry-After header when present.

Changed

• Default toolbar shortcut is now Ctrl+Shift+Y (Cmd+Shift+Y on macOS)
  instead of Ctrl+Shift+Space to avoid conflicts with input methods.
• Default Gemini model bumped from gemini-2.0-flash to gemini-2.5-flash
  (still on the free tier, GA, recommended by Google).
• Popup shortcut hint now reads the user's actual binding via
  chrome.commands.getAll, so customised shortcuts display correctly.