Voice Input

The chat panel supports dictating messages instead of typing them, using a local Whisper transcription pass. Nothing is sent to a cloud speech API — audio is captured and transcribed entirely on your machine.

Requirements

• Whisper — either the whisper Python package or the Whisper CLI tool available on your system PATH. BlenderMCP Pro checks for the Python package first, then falls back to checking for the CLI.
• PyAudio — required for microphone capture.

Neither is bundled with the addon; install them into Blender's Python environment (or a system Python that Blender's subprocess calls can reach) before using voice input. If neither is detected, the microphone button will report that voice input is unavailable rather than failing silently.

Using It

1. Click the microphone icon in the Chat tab.
2. Speak your request — recording runs for a set duration.
3. Whisper transcribes the recording locally once it completes.
4. The transcribed text is inserted into the chat input for you to review and send — it's not sent automatically, so you always get a chance to check the transcription before it reaches the AI.

Troubleshooting

• Microphone button shows unavailable — install PyAudio and either the whisper pip package or the Whisper CLI, then restart Blender.
• Transcription is slow — larger Whisper models are more accurate but slower. If you installed a large model, consider switching to a smaller one for faster turnaround during interactive work.
• No audio captured — check your OS-level microphone permissions for Blender; on macOS in particular, Blender needs explicit microphone access granted in System Settings.