GitHub Copilot in JetBrains: On-Device Voice Dictation (Preview)
GitHub Copilot in JetBrains IDEs now supports on-device voice dictation in public preview. Users can dictate chat messages on supported native plugin packages with Editor Preview enabled, after opting in and downloading a model. Transcripts stay as editable drafts until the user selects Send.
Key Takeaways
- On-device voice dictation is now in public preview for Copilot chat in JetBrains IDEs.
- It needs a supported native plugin package with Editor Preview enabled.
- Users opt in and download a model before dictating.
- Transcripts remain editable drafts until Send is selected.
- Dictation runs on the device rather than through a remote speech service.
- It is part of a broader JetBrains update that also added remote access and cloud sandbox.
Dictate Chat Messages in JetBrains
GitHub Copilot for JetBrains can now take voice input for chat messages. On-device voice dictation is in public preview and works on supported native plugin packages with Editor Preview enabled.
Getting Started
Users opt in and download a speech model to get started. Because the model runs on the device, dictation does not depend on sending audio to a remote service. Transcripts appear as editable drafts in the chat box and are not sent until the user selects Send, so mistakes can be corrected before Copilot sees the message.
The feature arrived in the November 17, 2026 JetBrains plugin update, which also added remote access, cloud sandbox and browser use controls.