The ChatGPT Photoshoot: Studio Headshots of Your Own Face (Higgsfield Plugin)
One plugin, a folder of selfies, and two prompts. You lock your own face as a reusable character, then hand ChatGPT any photo you like the look of and it shoots that photo with you in it.
What you need
Ten or so photos of your own face. Different angles, different lighting, no sunglasses, no heavy filters. Front on, three quarter, and profile if you have one. The model is learning your bone structure, so give it your face from more than one side.
Then the Higgsfield plugin inside ChatGPT.
⚡ Install the Higgsfield plugin for ChatGPT
Once it's installed you never leave the chat. The plugin gives ChatGPT the image models, so the whole shoot happens in one conversation.
Step 1: Lock your face as a character
Drop all your reference photos into ChatGPT in one message, then paste this. It builds a character lock with Sol 2.0, which is the thing that keeps your face identical across every image you generate afterwards.
This is the step everyone skips, and it's the one that matters. Without a lock you get a different stranger who vaguely resembles you in every shot.
Here are reference photos of me. Using Higgsfield Sol 2.0, build a character lock from these images.
Study my face across every photo. The bone structure, the jawline, the eye shape and spacing, the hairline, the skin tone and the skin texture. Build one consistent character I can reuse in every image from now on.
Do not average my features and do not beautify them. Keep my face exactly as it is, including anything a filter would normally smooth out.
Confirm when the character is saved and tell me what to call it so I can reference it later.Step 2: Build your look board
While the character is building, go to Pinterest and save the photos you actually want to look like. Search CEO headshot, editorial portrait, film photography portrait, or whatever fits what you're after.
Save six to ten. Pick shots that are genuinely different from each other, one clean corporate headshot, one moody side light, one outdoor golden hour, so you come out with a set instead of ten versions of the same picture.
What you are really choosing here is lighting and framing, not the person in the photo. A great reference is one where you could describe the lighting setup out loud.
Step 3: Shoot it
Drop your saved Pinterest photos into the same chat and paste this. It recreates each reference with your locked character in place of the model.
Boom. That's your headshots done, and you can keep feeding it new references forever because the character stays saved.
Here are reference photos of the look I want. Recreate each one using my locked character.
Match the reference exactly on lighting, camera angle, lens and depth of field, framing, wardrobe, colour grade and mood. The only thing that changes is the face, which is my locked character.
Shoot it as a real photograph, not a render. Keep the skin texture natural with visible pores and real shadow under the jaw. No plastic smoothing, no beauty filter, no waxy highlights.
Give me each shot at full resolution.The things that break it
Filtered reference photos. If your selfies are already smoothed by a beauty filter, the lock learns the filter and every output looks like plastic. Use the most unedited photos you have.
One angle only. Ten front facing selfies teach it a flat face. The moment you ask for a three quarter shot it invents the side of your head. Mix the angles.
References that are all the same. Ten corporate headshots gets you ten corporate headshots. Vary the lighting on your look board and the set does the work.
Not checking the first one. Generate a single image before you run the whole board. If the face is off, the lock is off, and it's cheaper to rebuild the character than to regenerate ten images.
Get the next one first
New prompts every week.
Free. The new drops and the tools behind them, before they hit the feed.
No spam · New issues Sunday · Unsubscribe anytime
Need it custom?
Want this built for you?
Tell me the idea and I’ll build it. An app, a tool, an automation. You don’t need to be technical.
Related guides
Prompt
Turn One Photo of Yourself Into a Cinema-Grade Anime (Claude + Higgsfield MCP)
The exact 3-step pipeline prompt that turns a single selfie into a consistent anime protagonist, a sidekick, and a 12-second cup-final scene, no drawing, no timeline.
Prompt
The AI Flash Filter: One Paste Turns Any Selfie Into a 2000s Party Candid
That flash-blasted disposable cam look all over your feed? Most of it never touched a camera. This is the prompt, word for word.
Prompt
Your Face in 30 Years: Healthy You vs Neglected You
One selfie in. ChatGPT ages your actual face 30 years, then puts the you who took care of yourself next to the you who didn't. Same face, two timelines.
Guide
The Higgsfield Prompt Vault: Every Workflow I Use, One Page
Every Higgsfield workflow from my videos in one place, the living website, the anime film, street interviews, drone flythroughs, plus every video prompt, copy-paste ready.
Frequently asked questions
Around ten, at mixed angles and in different lighting. Fewer than five and the character lock gets unreliable, especially at angles you never gave it.
You install the Higgsfield plugin into ChatGPT and work from there. Install it here, then everything happens inside your normal ChatGPT conversation.
The character lock either was not created or is not being referenced. Rebuild it with the prompt in step one, confirm ChatGPT names the character back to you, then call that name in your generation prompts.
Yes. The second prompt recreates whatever reference you feed it, so the same locked character works for editorial shots, travel photos, or fashion. Change the look board, not the prompt. Same idea as the anime version.
The giveaway is almost always over smoothed skin, which is why the prompt explicitly asks for visible pores and real shadow. If a result looks waxy, regenerate it rather than trying to fix it afterwards.