A channel is how you talk to your assistant. Right now, there's one primary channel with more on the way.
The main way to interact with your assistant. It's a native macOS application with:
The desktop app has full capabilities. Every tool, every skill, every feature is available here. This is the flagship experience.
You can talk to your assistant using your microphone. This works within the desktop app. Just click the microphone button or use the keyboard shortcut and start speaking.
Voice is great for:
Push-to-Talk: Hold your activation key and speak. The default is Right Option, but you can also use Fn, Right Command, or Right Control.
Wake word: Optionally enable hands-free activation by saying a keyword (e.g., “Hey Vellum”, “Computer”, or your assistant's name).
To set everything up, just say “Set up voice” — your assistant walks you through microphone permissions, key selection, wake word, and optional ElevenLabs text-to-speech, all in chat.
Not yet, but they're coming. The architecture is built to support multiple channels, which is why your assistant has its own identity and infrastructure separate from any single interface.
When new channels launch, your assistant will be the same assistant everywhere. Same personality, same memories, same skills. The only thing that changes is where you're talking to it.
💡 Why this matters: Some channels have fewer capabilities than others. A future SMS channel, for example, probably can't show you interactive apps. Your assistant is smart enough to adapt its output to what the channel supports. It'll give you a text summary instead of a dashboard, or a link instead of an embedded UI.