Built as an open-source voice powered alternative to Cursor for AI-powered codebase exploration and editing, Killclack extends the mission: to kickstart a movement away from keyboard-centric workflows, making code accessible through natural, spoken language. Our vision is to empower everyone—from seasoned developers to code newcomers—to create, command, and collaborate by voice.
We believe coding should be as natural as conversation.
Join the movement to kill the keyboard, starting with code!
Use KillClack just like you would VS Code for editing, navigating, and managing your codebase — but now with:
- Voice Commands for quick navigation, edits, and running tasks.
- Side Chat for asking explanations, generating code, or executing quick actions.
- Faster context switching without needing to move hands from the keyboard/mouse.
- The ability to combine typing + voice to work however is most efficient.
Think of it as VS Code + a fully integrated coding assistant that can hear you, understand your repo, and act instantly.
-
AI Voice Agents:
- Use industry-leading AI speech transcription via Deepgram
- Real-time voice-to-code dictation and code editing
- Voice-triggered commands: open files, search, run tasks, and more
- Fully hands-free coding and navigation, just by talking
-
Collaborative Voice Rooms:
- Join teammates in live, transcribed voice/video sessions (powered by Daily.co)
- See real-time transcript and participate in voice-based pair programming
-
AI Model Flexibility:
- Bring your own models or connect to any host/provider
- Messages go straight to providers—your data isn’t retained
-
Visualize and Checkpoint Changes:
- Visualize voice-driven edits and AI-generated code changes
- Traverse code edit history in new intuitive ways
-
Multi-language support for all coders, including new & emerging spoken languages
-
Welcoming the language of the people, not just the language of code!
See the INSTALL GUIDE for platform-specific setup (including Google Drive/GitHub repo ZIP download instructions and Docker requirements), or simply:
-
Install Killclack
- Download Killclack for your platform and install following the setup guide, or build from source.
-
Set Up Your API Keys & Voice Collaboration
- Go to the Settings or Voice section inside Killclack.
- LLM API Key: Enter your preferred LLM or AI provider’s API key if required (e.g. OpenAI, Azure, etc).
- Deepgram API Key: Enter your Deepgram API key.
Don’t have one? Get a free account at Deepgram Console. - Daily.co (Voice/Video Collaboration):
- Enter your Daily API key and paste a valid Daily room URL if you want to use collaborative rooms (for pair programming with live transcription & chat).
-
Open a Project
- Launch the Killclack app.
- Go to File → Open Folder... and select your coding repository (from GitHub, Google Drive, ZIP, etc).
-
Enable Your Microphone
- When prompted, allow Killclack (Void/VS Code) to access your microphone.
- If you’re not prompted, go to your Mac’s System Settings → Privacy & Security → Microphone and ensure Killclack is checked.
-
Connect Voice Agent (First-Time Only)
- Make sure Docker Desktop is installed and actively running before starting voice features.
- Killclack will automatically launch a secure Docker container for real-time speech and voice agent AI.
-
Start Voice Control
- You’ll see a a tab labeled "voice" in the bottom 'toggle bar'.
- If you dont see the bottom toggle panel with differnt tabs including terminal and voice, enable it by clicking the icon for bottom toggle panel (bottom bar) on the top right hand corner of the IDE. (Small rectagle on bottome of larger rectangle icon)
- Once you are on the voice tab, click the green telephone in the panel to connect to Kill Clacks Cody voice agent.
- Click to start listening, or simply use a wake word.
-
Use Wake Words & Hands-Free Voice Commands
- Hey there Cody — Wake up Killclack (activate voice control)
- CodyGo — Process your current request/command (“done with request”)
- CodySleep — Pause/sleep voice agent (stop listening)
-
Examples of What You Can Say
- “Open file README.md”
- “Search for TODOs”
- “Create a new Python file”
- “Run the build task”
- Or say code directly: “Print hello world in JavaScript”
- To join a Daily room: “Join the team room”
-
End/Restart Voice Mode
- Click the microphone icon to pause/resume listening.
- Say “CodySleep” to pause at any time; say “Hey there Cody” to reactivate.
- (Optional) Pair Programming & Voice Rooms
- With valid Daily credentials, join or invite teammates to live voice/code rooms with hands-free, transcribed collaboration.
- You can use all VS Code features alongside voice commands.
- Wake words = ultimate hands-free workflow.
- Privacy respected: all processing happens securely, using your keys.
- Voice not working?
Docker Desktop must be running! Check for the whale icon in your system tray/menu bar. - Mic not working?
Check your OS permissions:
System Settings → Privacy & Security → Microphone → enable Killclack. - No response to wake words?
Make sure the mic is active (icon is listening) and you speak clearly. - For logs, settings, key reconfiguration: Use Killclack’s settings menu.
Now you’re ready to code by voice—just say “Hey there Cody” and begin!
- 🧭 Website
- 👋 Discord
- 📄 HOW_TO_CONTRIBUTE
To get started, check out our Project Board and see HOW_TO_CONTRIBUTE. We run casual weekly meetings in our Discord—everyone’s welcome!
Killclack is all about expanding who (and how) people code. Whether you want to contribute code, docs, or ideas for new voice/AI workflows, we’d love your help.
Killclack is a fork of Void, which in turn is a fork of VS Code (original repo).
Work is temporarily paused on the original Void IDE (the parent repo) while the Void team experiments with novel AI coding concepts. Stay tuned for updates on the Void Discord.
- Join the movement: Discord link (https://discord.gg/nyFcbeqXeS)
- Email: [email protected]
Killclack: The future of code starts with your voice.
You can always reach us in our Discord server or contact us via email: [email protected].