Spoken progress updates
Hear when Cursor starts, finishes, hits errors, or needs input.
Cursor text to speech for voice-driven coding
Cursor text to speech that lets your AI assistant speak back.
TalkToCursor adds Cursor TTS through an MCP server, so your coding assistant can speak progress, completions, and errors while you work hands-free.
$ npx -y talktocursor
Why Cursor voice
Stop watching the chat window for every update.
Cursor can do long-running work, but you still have to keep checking the screen. Cursor text to speech turns those updates into spoken feedback while Cursor edits, debugs, explains, or asks for clarification.
Choose your speech engine
Use managed TalkToCursor Cloud, bring your own ElevenLabs key, or run Voicebox locally for free.
Speech controls
Choose minimal, brief, or detailed spoken responses, preview Cloud voices, and optionally pause playing media on macOS.
How it works
Add voice to Cursor without leaving your IDE.
TalkToCursor runs as a local MCP server. Cursor calls its speak tool during a task, and your selected Cloud, ElevenLabs, or Voicebox provider produces the audio. Speech is serialized to prevent overlapping updates; task-start speech can be non-blocking, and completion speech can hand off to Auto-Listen after playback.
1
Add the MCP server
Recommended: give the prompt below to Cursor or another coding agent. You can also install manually with npm.
Install TalkToCursor by following:
https://github.com/MindSyncTech/talk-to-cursor/blob/main/INSTALL.md
Configure the MCP and persistent voice-feedback instructions. Ask before enabling paid Cloud services or installing the macOS Background Helper.
https://github.com/MindSyncTech/talk-to-cursor/blob/main/INSTALL.md
Configure the MCP and persistent voice-feedback instructions. Ask before enabling paid Cloud services or installing the macOS Background Helper.
Restarts, sign-in, purchases, and macOS permissions may still require you.
2
Choose a TTS provider
Run npx -y --prefer-online --package talktocursor@latest talktocursor-settings, then connect TalkToCursor Cloud, add an ElevenLabs API key, or point Settings to a local Voicebox server.
3
Talk to Cursor
Use Cursor normally, or pair it with Wispr Flow for dictation.
Cursor on macOS only
Complete the hands-free workflow.
Use Wispr Flow or free local Handy for transcription. TalkToCursor’s Background Helper coordinates listening and submission without keeping a setup terminal open.
Smarter turn control, processed locally.
- Smart Turn v3.2 uses a verified 8.7 MB local model to detect when you finish speaking.
- Wake phrases can be required before a new instruction; improved sensitivity supports quieter speech.
- Say “send it” to submit immediately instead of waiting for turn detection.
- Check Permissions reports the helper’s macOS Accessibility and Microphone access.
1
Choose transcription
Install Wispr Flow or Handy, then select it in the hands-free section of Settings.
2
Configure turn behavior
Choose Fixed Pause or Smart Turn, then set wake phrase sensitivity and spoken “send it.”
3
Install the helper
Click Install & Start Background Helper, run Check Permissions, and grant the requested access.
Smart Turn, wake phrases, spoken submission, automatic submission, and the Background Helper are not claimed for Windows, Linux, Claude Code, Codex, Antigravity, or generic MCP hosts. See the full installation guide for setup and the manual fallback.
Search terms this solves
Cursor text to speech, without a separate app.
If you searched for Cursor TTS, Cursor text to speech, Cursor dictation, or Cursor voice, TalkToCursor is the direct install path. It keeps the workflow inside Cursor through MCP instead of forcing you into a separate coding interface.
- Text-to-speech for Cursor AI responses and task updates
- Voice presets for faster, slower, stable, expressive, or dramatic speech
- Optional auto-submit support for dictation workflows on macOS
- Settings UI for API keys, voice selection, and voice-loop controls
Recommended setup
Use Cursor TTS plus dictation for the full loop.
TalkToCursor handles the speaking side. For Cursor speech to text, use macOS dictation or Wispr Flow. For the full hands-free loop, pair TalkToCursor with Wispr Flow.
FAQ
Cursor TTS questions
Is Cursor TTS the same as TalkToCursor?
TalkToCursor is the product name. Cursor TTS describes what it adds: text-to-speech and voice workflow support for Cursor AI.
Does this replace Cursor?
No. It extends Cursor with an MCP server so your assistant can speak through TalkToCursor Cloud, ElevenLabs, or Voicebox while you keep working.
Do I need Wispr Flow?
No. TalkToCursor Cloud, ElevenLabs, or Voicebox handles text-to-speech. Wispr Flow or Handy is optional for speech-to-text and the hands-free loop.