You've installed Vellum. You've opened it. There's a friendly mascot on your screen waiting for you.
Now what?
Here's what your first 2 minutes look like. Think of it less like a tutorial and more like meeting someone new.
On first launch, you'll walk through a short setup:
Once your assistant hatches, it introduces itself with a simple message:
“Hey. I'm brand new, no name, no memories, nothing yet. The more we work together, the more context and memory I build, and the better I get. But let's not wait around. Throw a question at me, give me a task, or ask what I can do.”
There's no onboarding wizard or structured interview. Your assistant starts as a blank slate and learns about you organically as you interact. Give it a name, tell it about yourself, or just jump straight into a task — it's up to you.
Over time, it remembers your preferences, your projects, and how you like to work. The more you use it, the more useful it becomes.
Just start using it. Ask a question, give it a task, or tell it to set something up for you. There's no onboarding checklist to complete — your assistant is ready to work from the moment it hatches.
Try things like: “What can you do?”, “Set up my email”, or “Help me draft a message.” The best way to learn what it can do is to ask.
Your assistant always asks before taking action. When it needs to use a tool — like sending an email, running a shell command, or accessing a file — a permission widget appears directly in the chat. You choose how to respond:
If you're using voice mode, you can respond verbally: say “yes”, “no”, “10 minutes”, or “always”.
Nothing happens behind your back. You stay in control, and every sensitive action goes through you first.