Creative · Community report
Create an agent portrait brief
@c4chaos sent their Muse agent a TV clip of Muse's mascot, asked what it saw in a virtual mirror and got a portrait of their agent Fluffy, while @liuyuxxd turned a dog photo into an animated avatar through chat. This page shows how to write an agent portrait brief and review the image before using it.
What you can get
- Prepare
A character or agent description
- Ask and review
Pick a starting point: a question about how the agent sees itself, or a photo you own, such as your pet.
- Check the result
An approved portrait concept and, only if actually generated, a separately reviewed image artifact.
The source case
@c4chaos says they sent a television clip to their Muse agent, asked what it saw in a virtual mirror and received a portrait concept with cyan and magenta eyes, a jeweler’s loupe and circuit traces. The post describes the creative output; the captured text does not include the original image or establish which rendering tool produced it. It is a creative task, not evidence of machine consciousness. @liuyuxxd reports creating the agent's avatar through chat by submitting a photo of their dog and asking for a Disney-style treatment, which produced an animated avatar (named "Ario" by the user) with different animations for working and thinking modes; the post gives no detail on the image-generation tool or how many iterations it took.
A public report by someone outside this library. Not independently reproduced here.
From the source
Source excerpt
MEET FLUFFY: WHAT MY MUSE AGENT SEES IN THE VIRTUAL MIRROR
Source excerpt
10. Great design details and user experience. The design team really cooked. My biggest WOW moment is when editing the avatar. In an earlier version, I couldn’t find the flow in the UI. It turns out I can just create and tweak it through chat. eg I gave it a photo of my dog, asked it to turn to a disney style, and it created animation by default, and changed the other animations in working / thinking modes. Meet Ario here.
Quoted from the original posts. Not written by this guide, and not a result produced here.
Suggested prompt
Bring A character or agent description, Style references you have permission to use, Preferred output format and intended audience. Source community report, checked 2026-10-09.
Written by this guide as a starting point, not quoted from the source. Your edits stay in this browser.
What you need
You supply
- A character or agent description
- Style references you have permission to use
- Preferred output format and intended audience
How to try it
Editorial, not from the source
- Pick a starting point: a question about how the agent sees itself, or a photo you own, such as your pet.
- Ask for a written brief first, with each visual detail tied to something the agent does for you.
- Cut the brief to what will still read at avatar size: one face, two colours, one or two props.
- Generate one image, then view it at 48 pixels and at full size before asking for changes.
- If you want animated states, approve each one separately and check they match the still.
- Before using it publicly, check it does not resemble a known character, brand or real person.
Two ways the reports started
@c4chaos began with a clip: the Daily Show had put Meta's Muse mascot, Jolly, on air, and they sent that clip to their own agent and asked what it saw in the virtual mirror. They describe the result detail by detail: one cyan eye and one magenta, a jeweler's loupe on the forehead because the agent spends its days mining their fact-checks, circuit traces in the fur, and a creature called Noctopus in a background tank. "The only stakes in the portrait are mine," they write.
@liuyuxxd started from a photo. In a longer review of Muse, they say they could not find the avatar flow in the UI of an earlier version, then found they could create and tweak it through chat: they gave it a photo of their dog, asked for a Disney style, and it made an animation by default and changed the animations for working and thinking modes. They named it Ario. Neither post names the image tool or says how many attempts it took.
What a good portrait brief contains
The details in @c4chaos's description work because each one points at the agent's job: the loupe for fact-checking, the circuits for being software. A useful brief does the same. List three to five features, say what each stands for, fix two or three colours, and describe the style in plain words (flat, rounded, 3D, pixel) so you can adjust it later.
Plan for the size it will be seen at. Chat avatars are usually shown small and cropped to a circle, where a loupe or circuit pattern can turn into noise. Ask for a version with larger shapes and check it at small size before approving.
Before it goes public
A photo of your own pet is yours to use. A request in a named studio's style can drift towards that studio's characters, and a brand mascot belongs to the brand, so for anything public describe the features you want instead of naming the style. If you plan to use the image outside the app, check the image tool's terms for commercial use.
@c4chaos tags the post #AIConsciousness and plays with the idea of the agent seeing itself. This page treats the portrait as a creative task, not evidence of what the agent is.
Open Muse to try it. That is the product's own site; this guide is not part of it.
Limits and confirmation points
Limits to confirm
- Original portrait file, image-generation tool and complete exchange are not supplied.
- Visual details are reported by the author rather than verified from an image.
- This is an attributed community report, not reproduced by this guide. The suggested workflow and checks are editorial guidance, not a recovered conversation or guaranteed product capability.
If it goes wrong
Editorial, not from the source
- The portrait looks generic
- Add two or three details tied to the agent's actual work, and remove anything decorative that means nothing to you.
- Details disappear at avatar size
- Ask for fewer, larger shapes and stronger contrast, then check the image at 48 pixels in a circle crop.
- The result looks like a known character
- Drop studio or character names from the brief and describe shape, colour and texture instead.
Common questions
Can I change my Muse avatar through chat?
@liuyuxxd says they created and tweaked theirs through chat after not finding the flow in an earlier version's UI. This guide has not reproduced it, and the UI may have changed.
Can the avatar be animated?
@liuyuxxd says Muse made an animation by default and changed the animations for working and thinking modes. Ask to see each state before accepting it.
How do I make the portrait feel like my agent and not a stock image?
Tie each detail to what the agent does for you, as in @c4chaos's loupe for fact-checking, and keep the brief to a few details.
Which image tool does Muse use for this?
Neither report says, and neither image is included in the captured text.
Next
More in Creative
Similar tasks
Get recurring task recommendations — ask your own agent to check for relevant updates.