VS Code Speech
ExtensionFreeA VS Code extension to bring speech-to-text and other voice capabilities to VS Code.
Capabilities11 decomposed
voice-to-text chat input with hold-to-submit
Medium confidenceCaptures microphone audio during active chat sessions and transcribes it to text using Azure Speech SDK, with configurable language selection and automatic submission on release. Integrates directly into GitHub Copilot Chat UI via a microphone button, supporting both continuous listening and push-to-talk modes via Ctrl+I (Windows/Linux) or Cmd+I (macOS). The extension handles audio buffering, language detection, and real-time transcription without requiring API keys or internet connectivity for local processing.
Integrates Azure Speech SDK directly into VS Code's chat UI with hold-to-submit keybinding (Ctrl+I) rather than requiring separate voice recording apps or external transcription services; claims local processing without API keys, though Azure SDK dependency suggests potential cloud fallback architecture not fully transparent
Tighter VS Code integration than generic voice-to-text tools (Whisper, Google Speech-to-Text) because it's built into the editor's chat interface and respects VS Code's keybinding system, but lacks the offline-first guarantees of local Whisper models
editor dictation with cursor-position insertion
Medium confidenceEnables voice-to-text input directly into the active editor at the current cursor position via Ctrl+Alt+V (Windows/Linux) or Cmd+Alt+V (macOS). Uses Azure Speech SDK for transcription with configurable language selection. Text is inserted synchronously after transcription completes, supporting code comments, documentation, and prose without requiring chat context or Copilot Chat extension.
Operates independently of Copilot Chat, allowing voice dictation directly into any editor file without requiring AI chat context; uses VS Code's native keybinding system (Ctrl+Alt+V) and respects cursor position for precise insertion, unlike generic voice-to-text tools that require separate applications
More integrated than external dictation tools (Dragon NaturallySpeaking, OS-level speech input) because it's built into VS Code's editor context and respects cursor position, but lacks the AI-assisted correction and formatting of dedicated voice writing tools
development-stage extension with ongoing feature evolution
Medium confidenceThe extension is explicitly documented as 'still in development,' indicating active feature development, bug fixes, and potential breaking changes. The extension is distributed via the VS Code Marketplace as a free, installable extension, but stability, maturity, and feature completeness are not guaranteed. Users should expect changes to keybindings, settings, UI, and capabilities as the extension evolves.
Explicitly documented as 'still in development,' signaling that the extension is actively evolving and may undergo breaking changes; this transparency about maturity is rare among VS Code extensions, but creates uncertainty about long-term stability and feature completeness
More transparent about development status than many extensions that hide maturity issues, but less stable and feature-complete than mature voice tools (OS-native voice APIs, established voice platforms) that have reached production readiness
automatic text-to-speech synthesis of chat responses
Medium confidenceReads chat responses aloud using text-to-speech synthesis when the `accessibility.voice.autoSynthesize` setting is enabled AND the user initiated the chat message via voice input. The extension uses Azure Speech SDK for TTS with language selection matching the STT language. Audio playback occurs automatically after the AI response is generated, providing audio feedback without requiring manual activation.
Conditionally activates TTS only when STT was used as input (voice-in-voice-out pattern), rather than offering universal TTS for all chat responses; this reduces cognitive load and audio clutter for text-input users while providing full audio feedback for voice-first users
More contextually aware than generic TTS tools (OS-level screen readers, browser extensions) because it only synthesizes when voice input was used and integrates with Copilot Chat's response lifecycle, but lacks fine-grained control over voice selection and playback parameters
multi-language speech recognition and synthesis
Medium confidenceSupports speech-to-text and text-to-speech across 26 languages via the `accessibility.voice.speechLanguage` setting, which applies uniformly to both STT and TTS operations. Language selection is configurable via VS Code's Settings Editor and persists across sessions. The extension uses Azure Speech SDK's language models for both recognition and synthesis, with language detection and processing handled transparently without user intervention.
Provides unified language configuration (single `accessibility.voice.speechLanguage` setting) that applies to both STT and TTS, ensuring consistency across voice input/output workflows; leverages Azure Speech SDK's multilingual models rather than requiring separate language-specific tools
Broader language support (26 languages) than many open-source STT tools (Whisper supports ~99 languages but with variable quality), but less granular than enterprise speech platforms (Google Cloud Speech-to-Text, AWS Transcribe) which offer per-request language selection and custom vocabulary
keybinding-driven voice session control
Medium confidenceProvides keyboard shortcuts to start, stop, and submit voice input sessions without mouse interaction. Default keybindings are Ctrl+I (Windows/Linux) or Cmd+I (macOS) for chat voice (hold-to-submit or toggle mode), and Ctrl+Alt+V (Windows/Linux) or Cmd+Alt+V (macOS) for editor dictation. Keybindings are fully customizable via VS Code's Keybinding Shortcuts Editor, with conditional activation via `when` clauses (e.g., `!voiceChatInProgress`, `!editorDictation.inProgress`) to prevent conflicts.
Integrates with VS Code's native keybinding system and `when` clause conditions, allowing voice session control to be composed with other editor state checks (e.g., `when: editorFocus && !voiceChatInProgress`); supports both toggle and hold-to-submit modes via keybinding configuration
More flexible than fixed voice activation buttons (Copilot Chat's microphone icon) because it respects VS Code's keybinding customization system and conditional activation, but requires manual configuration compared to out-of-the-box voice tools with preset keybindings
local speech processing with azure speech sdk
Medium confidenceProcesses speech-to-text and text-to-speech operations using Azure Speech SDK, which the extension claims performs local processing on the user's machine without requiring internet connectivity or API keys. The SDK handles audio capture, buffering, language detection, and transcription/synthesis internally. However, the documentation does not explicitly clarify whether Azure Speech SDK calls are truly local or cloud-based, creating ambiguity about data privacy and network requirements.
Claims local speech processing via Azure Speech SDK without requiring API keys or internet connectivity, positioning as a privacy-first alternative to cloud-based STT/TTS services; however, the actual architecture (local vs. cloud) is not transparently documented, creating uncertainty about data handling
Avoids the API key management and cloud service costs of Google Speech-to-Text or AWS Transcribe, but lacks the transparency and offline-first guarantees of local Whisper models; Azure Speech SDK's true processing location (local vs. cloud) is ambiguous compared to clearly local alternatives
github copilot chat ui integration with microphone button
Medium confidenceEmbeds a microphone button directly into the GitHub Copilot Chat interface, providing visual affordance for voice input without requiring keybinding knowledge. The button appears in the chat input area and triggers voice capture when clicked or held, with visual feedback indicating recording state. Integration is seamless when both VS Code Speech and GitHub Copilot Chat extensions are installed; the microphone button is unavailable if Copilot Chat is not present.
Provides native UI integration with GitHub Copilot Chat's chat input area via a microphone button, rather than requiring users to discover and memorize keybindings; the button is context-aware and only appears when Copilot Chat is available, avoiding UI clutter
More discoverable than keybinding-only voice input (Copilot Chat's default) because the microphone button provides visual affordance, but less flexible than keybinding-driven activation because it's limited to Copilot Chat and cannot be customized or extended to other chat interfaces
cross-platform voice support with os-specific permission handling
Medium confidenceProvides voice capabilities across Windows (x64/ARM), macOS (x64/ARM), and Linux (x86/x64/ARM32/ARM64) with platform-specific microphone permission handling. On macOS, users must explicitly grant microphone access via Privacy & Security settings; on Windows and Linux, permission mechanisms are not documented. Linux support requires ALSA shared library (libasound) installation. The extension abstracts platform differences via Azure Speech SDK, presenting a unified voice API across all platforms.
Abstracts platform-specific microphone permission handling via Azure Speech SDK, supporting both x64 and ARM architectures across Windows, macOS, and Linux; Linux support requires explicit ALSA library installation, making it more complex than macOS/Windows but more flexible than platform-specific voice tools
Broader platform support (Windows, macOS, Linux with ARM variants) than many voice tools that focus on macOS or Windows only, but requires more manual setup on Linux (ALSA library) compared to OS-native voice APIs (Windows SAPI, macOS AVFoundation)
voice session state management with conditional keybindings
Medium confidenceTracks voice session state (active/inactive) for both chat voice and editor dictation, exposing state via `when` clause conditions (`voiceChatInProgress`, `editorDictation.inProgress`) that can be used in keybindings to prevent conflicts or trigger conditional actions. The extension manages state transitions (start, recording, stop, submit) internally and prevents simultaneous voice sessions across chat and editor contexts.
Exposes voice session state via VS Code's `when` clause system, allowing keybindings to conditionally activate based on voice recording status; this prevents conflicts and enables sophisticated workflows, but lacks a programmatic API for extensions to react to state changes
More integrated with VS Code's keybinding system than external voice tools, but less flexible than a full event-driven API because state is only accessible via `when` clauses and not exposed to other extensions
freemium licensing with free core voice features
Medium confidenceOffers voice-to-text and text-to-speech capabilities at no cost via the free tier, with no documented premium tier or paid features. The extension is distributed via the VS Code Marketplace as a free, open-to-install extension with no license key, subscription, or payment requirement. Pricing model is freemium, but the premium tier (if any) is not documented.
Provides core voice capabilities (STT, TTS, chat integration, editor dictation) at no cost via the free tier, with no documented premium tier or paid features; this contrasts with many voice tools that require API keys, cloud service subscriptions, or premium licenses
More accessible than paid voice tools (Google Cloud Speech-to-Text, AWS Transcribe, specialized voice editing software) because it's free and built into VS Code, but lacks the advanced features, customization, and support of enterprise voice platforms
Capabilities are decomposed by AI analysis. Each maps to specific user intents and improves with match feedback.
Related Artifactssharing capabilities
Artifacts that share capabilities with VS Code Speech, ranked by overlap. Discovered automatically through the match graph.
Speechnotes
Your Efficient Speech-to-Text...
Wispr Flow
Flow makes writing quick with seamless voice dictation for any application on your computer.
ChatGPT Writer
Revolutionize writing: AI-enhanced emails, error correction, tone adjustment, multilingual...
RealChar
Audio-driven interactions, users can record their voice to generate lifelike responses from AI-generated...
Cleft
Transforms voice to structured markdown notes, ensuring privacy and...
Dictation IO
Transform speech into text instantly, enhancing productivity across...
Best For
- ✓developers with accessibility needs (RSI, mobility constraints)
- ✓solo developers seeking faster code exploration via voice queries
- ✓teams using GitHub Copilot Chat as primary AI assistant
- ✓technical writers documenting code via voice
- ✓developers seeking faster documentation generation
- ✓early adopters willing to tolerate breaking changes and bugs
- ✓developers seeking to provide feedback on voice features
- ✓teams evaluating voice capabilities before they reach stable release
Known Limitations
- ⚠Requires GitHub Copilot Chat extension installed; chat voice features unavailable without it
- ⚠Language support limited to 26 languages (specific list not enumerated in documentation)
- ⚠No multi-turn voice conversation without manual re-triggering between exchanges
- ⚠Transcription accuracy depends on microphone quality and ambient noise; no noise cancellation documented
- ⚠Azure Speech SDK dependency suggests potential cloud fallback behavior not explicitly documented
- ⚠Insertion point fixed to current cursor position; no multi-location or batch insertion
Requirements
Input / Output
UnfragileRank
UnfragileRank is computed from adoption signals, documentation quality, ecosystem connectivity, match graph feedback, and freshness. No artifact can pay for a higher rank.
About
A VS Code extension to bring speech-to-text and other voice capabilities to VS Code.
Categories
Alternatives to VS Code Speech
Are you the builder of VS Code Speech?
Claim this artifact to get a verified badge, access match analytics, see which intents users search for, and manage your listing.
Get the weekly brief
New tools, rising stars, and what's actually worth your time. No spam.
Data Sources
Looking for something else?
Search →