Aar318's Twitch TTS

Text-to-Speech for Twitch chat with resilient playback and natural username pronunciation.

Aar318's Twitch TTS: Customizable Text-to-Speech for Twitch Chat Listen to Twitch chat while you watch. Twitch TTS reads new chat messages aloud with your browser's built-in speech voices and gives you direct control over the voice, volume, usernames, bots, repeated emoji, and the on-page controls. New Features: - Added compact, touch-friendly controls for phone-width Twitch pages. - Improved the mobile settings panel to fit the visible screen and device safe areas. - Prevented horizontal overflow and unwanted form zoom on narrow screens. - Improved settings accessibility labels and dialog behavior. - Typing in Twitch chat no longer disrupts the TTS queue. Unexpected browser pauses are resumed, and interrupted messages continue near the last spoken word. - Added optional smart username pronunciation for stylized names such as Cl0V3R_o7 (Clover07) and R430r (Razor). - Custom username mappings still take priority over automatic pronunciation. - Combined TTS and Settings into one clean segmented control with independent click targets. - Fixed a queue-stall edge case after the enable announcement or voice preview. - The active TTS segment now has a persistent green border. - Editing custom name mappings no longer restarts active speech; recovery waits for typing to settle and advances beyond the completed word boundary. - First-session emoji-only messages announce the chatter and emoji description. - Added announcements for native Twitch Cheers/Bits and recognizable StreamElements/Streamlabs-style tips. - Expanded extraction for image emotes and messages without ordinary text fragments. - All settings, mappings, and allowed-bot entries now auto-save; complete settings export/import is available in the panel. - Enable/disable feedback now uses priority retryable announcements, and first-session emoji-only messages explicitly say “emoji” after the chatter's name. - Prevented cold-start clipping of “TTS” in Enabled/Disabled announcements and corrected interruption recovery so a username is never treated as spoken before its word completes. - Twitch username tags are now spoken as “at [name],” irregular mixed-case names are normalized more naturally, and Settings Backup controls now appear at the top of the panel. - Fixed Twitch's split mention markup so tagged messages are spoken completely, including all text after the @username. - Message lines now wait briefly for all mention and text fragments to settle before entering the speech queue. - Removed the hard-coded personal pronunciation mappings; only user-created custom mappings override smart username pronunciation. - The editable custom-mapping field is now seeded once with `hor318:Dev` and `xthexpanx:Spanks`; both entries can be changed or cleared normally without returning. - Settings import now offers the Stream Notifier-style “Keep existing settings (merge)” choice. Merge preserves current preferences and conflicting mappings while adding missing mappings and allowed bots; unchecked import overwrites the backup scope. Key Features: Real-Time Chat Reading: Watches Twitch chat and queues new messages for speech without requiring a separate account or external text-to-speech service. Browser Voice Selection: Choose from the speech voices installed or exposed by your browser and operating system, adjust volume from 0 to 100 percent, and preview the selected voice. Easy Controls: Toggle TTS from the injected Twitch button or by left-clicking the extension icon. Right-click the icon to toggle TTS, open settings, stop speech immediately, or open support links. Speak Selected Text: Select text on a Twitch page, right-click it, and choose "Speak selected text" to hear it through the current voice and volume settings. Custom Name Pronunciations: Map Twitch usernames to the name or pronunciation you want spoken. Mappings accept JSON or simple username:value lines. Username Autocomplete: Recently observed chat usernames are cached locally and offered while editing name and bot lists, making custom entries faster to create. Bot Controls: Common bots are excluded by default. Optionally enable bot responses and provide an allowlist of bots that should still be read. Emoji and Command Cleanup: Compress repeated emoji into phrases such as "emoji times 3," choose the minimum repetition threshold, and make exclamation commands easier to understand when spoken. Queue Protection: Suppresses recently duplicated messages and limits the speech queue so very busy chats do not grow without bound. Holiday Themes: Optional seasonal styling and emoji for the TTS controls, with a settings preview showing the current and next holiday. Optional On-Page Controls: Hide or show the injected TTS button while retaining toolbar and right-click controls. Privacy Focused: Settings and autocomplete data remain in local browser storage. Chat is processed on the page and spoken through the browser's Web Speech API; no chat text is sent to an external TTS server. This extension works only on Twitch.tv and starts disabled by default. Permissions Usage: Storage: Saves the enabled state, selected voice, volume, custom name mappings, bot settings, emoji preferences, holiday-theme preference, injected-button visibility, and recent-username autocomplete cache. Context Menus: Adds extension-icon commands for TTS controls and a user-invoked "Speak selected text" command on Twitch pages. Active Tab: Lets extension-icon and context-menu actions communicate with the content script on the current Twitch tab. Host Permissions: Limited to https://www.twitch.tv/* so the extension can observe Twitch chat and add its controls.