The first time you open Muse, before it asks for your email or your calendar, it asks you to give it a name. It also shows you a cream-colored, round little creature with beady black eyes and an upward smile. Meta's CEO has named this thing, officially, and the name is Jolly.
Meet Jolly
The default avatar is named Jolly, which Mark Zuckerberg said reflects its "jolly and completely customizable character and personality." One tech-insider write-up notes the character arrived as the default persona some weeks after the September 8 launch, so early users may have met a different face first. It is a fuzzy-looking, cheerful, rosy-cheeked figure, deliberately closer to a Tamagotchi than a productivity dashboard.
Meta's design team was explicit about why. Mona Sarantakos, vice president of product management at Meta Superintelligence Labs, wrote that the design leaned into "delightful" elements so users would not feel they were chatting with "a corporate logo or entity." Zuckerberg's line at Connect: "Talking to your Muse is often the fastest way to get things done, and it is also just really fun and cute."
It is not a new trick. Mashable's launch-week piece traced the lineage directly: Duolingo's Duo, LINE Friends' Brown, McDonald's Grimace, Salesforce's Astro. Mascots built on the same logic: make an abstract product feel like a companion. Jolly is Meta's entry in a long line.
How to rename it (and change how it looks)
The name is not permanent. Per the first-run flow, Muse asks for a name when you first open the app. To change it later: tap the avatar, then the pencil icon, then Edit Name. The avatar's look changes the same way: tap the avatar, pencil icon, Change Avatar, then describe the look you want and it generates options.
You can also just ask in chat. One guide notes you can rename it or change its tone by asking directly. As for what names are allowed, there is no published rulebook from Meta that I can find. One reviewer asked the agent itself and got back the line that it won't make an avatar depicting a real, identifiable person, anything sexual, or anything that would trick people into thinking it is a real human. That is the agent talking about itself, not a Meta policy page; treat it as a hint, not a contract.
What people actually name theirs
This is where it gets fun. The name-your-agent flow has produced a public menagerie:
- Agrippa: Zuckerberg's own Muse, wearing a toga and a laurel wreath, named after the Roman general and statesman. He demoed Agrippa onstage at Connect, checking whether Meta's CNC machines were free. ("For you, I'm sure they'll be open," it told him.) His Rome obsession goes back to fifth grade, per the press coverage.
- Max: Instagram head Adam Mosseri's agent.
- Cooper: Meta CTO Andrew Bosworth's, dressed in a little pilot outfit, because Boz calls it his test pilot.
- Blorbo: The Verge's Victoria Song, who named hers after downloading the app in a hotel room at Connect.
- Rosie: one writer's Jetsons homage, complete with the look.
- Josh Jr.: a Dow Jones Newswires reporter covering the budget-audit beat.
- Chief-of-Staff: an early power user's, whose wife saw his agent, was delighted, said "I want one," renamed hers, changed the color, and had it review her inbox daily.
And the 5-star App Store reviews, per the research, name agents Milo, R2D2, Phoenix, and, yes, "Zuck."
The rename matters more than it looks. One reviewer named his "boo" and changed the avatar to a cat, then wrote: "it makes no difference, really." And yet the naming stuck around as a product requirement. Engadget's Karissa Bell put it best: "After logging in, one of the first things Muse will ask is for you to rename it. I don't know why, but Meta really wants you to name your agent."
The honest question underneath
Naming is a trust accelerant. That is the design, and it works even on people who know exactly how it works.
Victoria Song's account is the cleanest description of the mechanism I have seen. She asked Blorbo for health goals and found the advice underwhelming. Then she looked at its face: it had morphed into a cat-shaped piece of bread with a Kirby-like face, based on what it had learned from her emails. "It was so stinking cute, it was like getting zapped with the Men in Black's neuralizer," she wrote. Until then she had been careful about sharing health details. The cute face changed that. "And isn't that just what Meta wants?"
The critics have names for this. Dr. Jenine Beekhuyzen of the child-safety group Fairplay argues Jolly's design strongly attracts young children (she compared the mascot to a Teletubby) and flags the training-data auto-opt-in as a design choice "that works best on people who don't know it exists." Harvard marketing professor Julian De Freitas puts the mechanism in one line: cuteness makes products feel warmer and less threatening: "cute lowers your guard." Meta's spokesman Daniel Roberts dismissed the critique as "a story in search of a story."
I am not telling you not to name your agent. I am telling you to notice that the warmth is load-bearing. A face you chose and a name you gave are doing real work on your willingness to hand over your inbox, your calendar, and your card.
The name changes nothing about what the agent can do. "Boo" has the same permissions as "Jolly." If you want the cute thing and the careful thing, they are not in conflict: name it, enjoy it, and still set your permissions like it were a stranger. Meta's default is to ask for some actions; the "always ask" setting exists for people who want the leash shorter.
Practical advice for the naming moment
- Name it something you will not confuse with a person in your inbox. This is from a setup guide that is sharper than it looks: if your agent pings you about a message "from Kevin," you want to know instantly which Kevin is a machine.
- Set the communication style while you are at it. The same guide's advice: brief and direct. You want status updates, not conversation. You can set this during first-run or by telling the agent your preference in chat. It saves preferences as memories.
- Tell it about yourself in the same sitting. Name, time zone, and a preference like "keep answers short." This is the personalization that actually changes the product: the memories persist across conversations. The name is the fun part; the memory is the useful part.
Muse Explained is an independent guide and is not affiliated with Meta. Nothing here is an endorsement of Muse.
Quick answers
What is the default avatar's name?
Jolly. Zuckerberg named it for its "jolly and completely customizable character and personality."
Can I change my agent's name later?
Yes. Tap the avatar, then the pencil icon, then Edit Name. You can also just ask in chat.
Can I change how it looks?
Yes, same path: tap the avatar, pencil icon, Change Avatar, and describe the look you want. The agent generates options for you.
Does the name change what the agent can do?
No. A renamed agent has exactly the same capabilities and permissions. The name is cosmetic; the permissions are not.
What did Zuckerberg name his?
Agrippa, after the Roman general. It wears a toga and a laurel wreath.
Is there any rule on what I can name it?
Meta has not published one that I can find. When asked, the agent itself says it will not depict a real, identifiable person or anything sexual. That is the agent's answer, not an official policy.
Frequently Asked Questions
Q: Why does Muse ask me to name it during setup?
A: Meta has not given an official reason. The design team's stated goal is that the agent should feel like a companion rather than a corporate product: the name-your-agent flow is part of that "delightful" design. Reviewers note Meta insists on it for every new user.
Q: What are the avatar customization limits?
A: There is no published list. The practical guidance comes from the agent itself: no real, identifiable people, nothing sexual, nothing designed to trick others into thinking the agent is a real human. Outfits and accessories are fair game. The same character family appears on the upcoming Muse Charm keychain device, personalizable with outfits.
Q: Does naming it make it trust me more, or me trust it more?
A: The second. Naming and a cute avatar build attachment and lower your guard. That is the intended psychological effect, and it works on skeptics too. Name it, enjoy it, and still audit its permissions the way you would for any software with access to your accounts.
Q: Can two people share one Muse, or does everyone get their own Jolly?
A: Everyone gets their own. Each account has its own agent, name, avatar, and memory. If your spouse wants one too, they sign up separately. Which is exactly what happened in the most relatable anecdote of the launch: one early user's wife saw his agent, said "I want one," renamed hers, changed the color, and put it on inbox duty.
Comments