Changelog

Unreleased

• Replaced the name-prefix pitch overrides with leader pitch: whatever the party or raid leader says is read at max pitch, and it follows the role so it moves with leadership instead of matching a fixed name.
• Settings panel under Interface > AddOns > Group Chat TTS for the on/off toggles and sliders, so you don't have to remember the slash commands. Open it with /gct options. The commands still work and stay in sync with the panel.
• Split the single GroupChatTTS.lua into Core/, Chat/, and UI/ files. No behaviour change; the addon table moved from file-locals onto the shared GCT table.

0.1.0

• Initial release
• Reads party/raid/leader/raid-warning chat aloud via C_VoiceChat TTS
• Assigns each group member a random, preferably-unique voice on join
• Announces the assigned voice in chat + spoken sample
• Queued playback, chat-escape cleanup, /gct settings command
• /gct config — movable window to view/change each member's voice (with preview)
• "Yell" mode: raid warnings + ALL-CAPS/!! messages read with raised pitch & max volume via SAPI XML prosody (/gct yell, /gct yellpitch)
• /yell channel support: reads nearby players' yells (incl. your own), each with a random voice from a separate non-pruned pool (/gct readyell)
• Language filter for random assignment (/gct langs; default English + Finnish); manual /gct config picks still allow any voice. Languages with no offline voice (Finnish) keep their Online voices despite excludeOnline.
• Pitch-offset fallback: when distinct voices run out, reused voices get a distinct pitch so people still sound different (uniqueness over the voice+pitch pair)
• Name-prefix pitch overrides (namePitchOverrides; default "Ale" -> max pitch)
• Configure your own voice in /gct config (persists; used when readSelf/readYell on)
• Mute while in combat, resume after (/gct combat, default on); also resets the speech queue on combat start/end so it can't get stuck silent
• Markup test lab (/gct lab): try each markup feature per voice with sliders/dropdowns, record works/fails per voice (persisted), and /gct dumptests to print results