Overview
Voice Mode in the SynapseX VS Code Agent enables a hands-free, conversational coding experience powered by ElevenLabs speech technology. Speak your queries naturally, and SynapseX will listen, understand, and respond both in text and audio — all within VS Code. You can use Voice Mode for any task you’d normally type: generating code, asking questions, requesting edits, debugging, or planning architecture — simply by speaking.Voice Mode is powered by ElevenLabs speech-to-text and text-to-speech technology.
Ensure your microphone is enabled and accessible by VS Code before using this feature.
Enabling Voice Mode
Open the SynapseX Chat Panel
Click the SynapseX icon in the VS Code Activity Bar to open the chat panel.
Locate the Voice Button
Find the microphone icon in the bottom toolbar of the chat panel (between the
Add Context button and the Camera button).
Using Voice Mode
Starting a Conversation
Once Voice Mode is enabled, simply begin speaking your query. SynapseX will transcribe your speech in real time and display the text in the chat input field. Speak naturally — SynapseX handles punctuation and code terminology automatically. Example voice queries:- “Add error handling to the login function”
- “What’s the best way to implement JWT refresh tokens in this project?”
- “Explain what this selected code does”
- “Create a React component for a user profile card with TypeScript”
Interacting with Responses
SynapseX responds with both text in the chat panel and an audio response read by ElevenLabs. You can:- Read the full response in the chat panel
- Listen to the audio summary while you keep your hands on the keyboard
- Follow up with additional voice queries seamlessly
Muting and Unmuting
To pause microphone input without ending the session:- Mute: Click the microphone icon once — the icon changes to indicate it is muted
- Unmute: Click the microphone icon again to resume listening
- The text response is always visible in the chat even while muted
Capabilities
Code Analysis
Ask SynapseX to explain code, review logic, or analyze your entire codebase vocally.
Code Editing
Request specific edits, refactors, or additions by speaking naturally.
Debugging
Describe errors out loud and get step-by-step debugging guidance with audio feedback.
Feature Development
Plan and build new features end-to-end through a natural voice conversation.
Repository Insights
Ask high-level questions about your project structure, dependencies, or architecture.
General Assistance
Get answers to documentation questions, programming concepts, or design decisions.
Example Voice Queries
Requesting Code Edits
“Add the capability to handle multiple concurrent user requests to the API handler”SynapseX will analyze the relevant files, generate the implementation, and apply the changes while verbally explaining what was done.
Summarizing Your Repository
“What databases are being used in this project and how are they connected?”SynapseX will scan your codebase and provide an audio and text overview of your data layer.
Debugging an Error
“Why is the payment API returning 500 errors on checkout?”SynapseX will trace the error through your code and suggest fixes, reading out the most likely cause and solution.
High-Level Planning
“Design a new user authentication system using OAuth 2.0 and JWT refresh tokens”SynapseX will outline a complete architecture plan and walk you through implementation steps in natural speech.
Best Practices
Speak Clearly and Specifically
Speak Clearly and Specifically
Use specific technical terminology when possible. Instead of “fix the thing”,
say “fix the null pointer exception in the UserService class”.
Confirm Complex Requests
Confirm Complex Requests
For multi-step tasks, confirm each step before proceeding:
“First, create the database schema. Does that look right?”
Mute When Not Speaking
Mute When Not Speaking
Mute the microphone when taking calls, playing audio, or when background noise
is high to prevent accidental input.
Review Text Transcripts
Review Text Transcripts
Always check the text transcript in the chat panel to ensure speech was
correctly recognized, especially for variable names, file paths, or complex terms.
Combine with Visual Context
Combine with Visual Context
Select a code section or take a screenshot before using Voice Mode to give
SynapseX precise context for your spoken query.
Ending a Voice Session
To end the Voice Mode session:- Click the end session (hang-up) button that appears in the toolbar while Voice Mode is active
- The microphone deactivates and the audio session closes
- The complete conversation history remains in the chat panel for reference
Summary
Voice Mode transforms SynapseX into a hands-free AI coding partner. With ElevenLabs powering both speech recognition and audio responses, you can stay in flow, keep your hands on the keyboard, and communicate with your AI assistant as naturally as speaking with a colleague.Key Features
Explore all VS Code Agent features
Best Practices
Get the most from SynapseX