Audio Cursor for VS Code

Audio Cursor brings intelligent text-to-speech reading with synchronized editor word highlighting directly into your VS Code environment.

Whether you are proofreading code comments, reviewing documentation, listening to terminal output, or listening to code aloud for accessibility, Audio Cursor lets you listen comfortably with both high-fidelity natural AI voices and offline system voices.

Features

1. Two Convenient Surfaces

Status Bar Item: Displays play/pause status, speech progress percentage (42%), estimated listening time, and word count. One click starts or pauses reading.
Activity Bar & Sidebar Player: Full-featured player with scrub bar, interactive text preview, sentence navigation, preset speed pills (0.8×–2.0×), and voice selector with language/gender/type filter dropdowns.

2. Synchronized Word Highlighting

The currently spoken word is highlighted in real-time in the active editor.
Follow Cursor / Auto-Scroll: Keeps the spoken word in view without fighting manual scrolls.

3. Reads Any Text in VS Code — Including Terminal & Markdown

Code & Documents: Select any block of text in any editor — files, untitled buffers, diffs, output channels, notebook cells, and remote documents.
Terminal: Select output in an integrated terminal and press Alt+P or click Read terminal selection in the sidebar player.
Markdown as Prose: Reads markdown files and preview tabs naturally without reading raw syntax symbols.
Cursor-to-EOF: If nothing is selected, Alt+P starts reading smoothly from the cursor position to the end of the file.

4. 340+ Natural AI & Offline Voices

Browse and search over 340 natural AI neural voices across dozens of languages and accents, plus all offline system voices.
Filter by Language, Gender, and Engine Type using the compact filter dropdowns.

5. Interactive Text Preview & Seeking

Click any word in the sidebar text pane or drag the scrub bar to jump playback instantly to that exact position.
Navigate sentence-by-sentence using Previous Sentence (Alt+[) and Next Sentence (Alt+]) controls.

6. Customizable Speech Settings

Adjust playback speed (0.5x – 2.0x) with quick preset pills (0.8x, 1.0x, 1.2x, 1.5x, 2.0x) and fine-tune voice pitch.
Native VS Code theme matching using official theme tokens (--vscode-*).

Keyboard Shortcuts

Shortcut	Command	When
`Alt+P`	Toggle Play / Pause	While focused in an editor, or during playback
`Alt+P`	Read Terminal Selection	While focused in an integrated terminal
`Alt+P`	Toggle Play / Pause	While a Markdown preview is focused
`Alt+Shift+P`	Stop	While playing or paused, from any surface

Previous / Next Sentence ship without a default shortcut — bind them yourself from Audio Cursor: Configure Keyboard Shortcuts, or the Remap… button in the sidebar.

Shortcuts pressed inside a terminal only reach the extension if the command is listed in terminal.integrated.commandsToSkipShell. Audio Cursor contributes its own, so this works out of the box; if you have set that value yourself it replaces the contribution, and Audio Cursor will offer to add the missing entries back.

Commands

All commands can be found in the Command Palette (Ctrl+Shift+P / Cmd+Shift+P):

Audio Cursor: Play/Pause (audioCursor.togglePlayback)
Audio Cursor: Play (audioCursor.play)
Audio Cursor: Pause (audioCursor.pause)
Audio Cursor: Stop (audioCursor.stop)
Audio Cursor: Read Selection (audioCursor.readSelection)
Audio Cursor: Read Terminal Selection (audioCursor.readTerminalSelection)
Audio Cursor: Read From Cursor (audioCursor.readFromCursor)
Audio Cursor: Next Sentence (audioCursor.nextSentence)
Audio Cursor: Previous Sentence (audioCursor.previousSentence)
Audio Cursor: Select Voice (audioCursor.selectVoice)
Audio Cursor: Speed Up (audioCursor.increaseRate)
Audio Cursor: Slow Down (audioCursor.decreaseRate)
Audio Cursor: Open Player (audioCursor.openPanel)
Audio Cursor: Show Logs (audioCursor.showLogs)

Configuration Settings

Configure Audio Cursor via Settings (Ctrl+, / Cmd+,) under audioCursor:

Setting	Type	Default	Description
`audioCursor.hostAudio`	boolean	`true`	Play audio from the extension host (Windows) so the first read of a window needs no click in the panel. Natural Neural voices only.
`audioCursor.voice`	string	`"en-US-JennyNeural"`	Preferred TTS voice. A Natural Neural voice name (needs a network connection), one of your operating system's offline voice names, or `system` for the OS default voice.
`audioCursor.rate`	number	`1.0`	Speech rate multiplier (0.5 to 2.0).
`audioCursor.pitch`	number	`1.0`	Speech pitch multiplier (0.0 to 2.0).
`audioCursor.highlightWord`	boolean	`true`	Highlight the currently spoken word in the active editor.
`audioCursor.followCursor`	boolean	`true`	Automatically scroll editor to keep spoken word in viewport.
`audioCursor.readFromCursorWhenNoSelection`	boolean	`true`	When no text is selected, read from cursor to EOF.
`audioCursor.statusBar`	string	`"auto"`	Status bar visibility: `auto`, `always`, or `never`.
`audioCursor.autoRevealPanel`	boolean	`true`	Reveal the sidebar player the first time playback needs to start the audio engine. Never takes keyboard focus, and never fires again once the player is initialized.
`audioCursor.stopOnDocumentChange`	boolean	`false`	Stop playback if document is modified during reading.
`audioCursor.readMarkdownAsProse`	boolean	`true`	Speak Markdown as prose without reciting markdown syntax.
`audioCursor.chunkSize`	number	`300`	Target character length per speech chunk.
`audioCursor.queueAhead`	number	`12`	Number of chunks queued ahead in speech engine.
`audioCursor.sanitizeCode`	boolean	`false`	Clean up excessive code punctuation before speaking.

Architecture & Technology

Audio Cursor uses a zero-runtime-dependency architecture:

VS Code Extension Host: Manages document selection snapshots, token chunking, session lifecycle, status bar updates, and editor decorations.
Sidebar Webview: Runs the audio engine with real-time neural WebSocket audio streaming and Web Speech API fallback, delivering low-latency audio output with word-level boundary synchronization.

Audio Cursor

ProdigyCorp