Build a Fully Interactive 3D Model Inside ChatGPT (GPT-6 Astra + Higgsfield)
GPT-6 Astra plans and builds. The Higgsfield plugin handles the visuals. Put them in the same ChatGPT conversation and you can ask for a working interactive 3D tool and actually get one back. The exact one from the video is linked below, so you can click around it before you build your own.
See the one from the video
This is the finished build, live. Open it and click something.
🫀 Open FORMA, the interactive anatomy atlas
Built start to finish in one ChatGPT conversation with GPT-6 Astra and the Higgsfield plugin. No 3D software, no modelling, no code editor.
2,234 separate anatomical pieces across 15 systems and 3,432 named structures. Drag to orbit, scroll to zoom, click any single piece to isolate it and read what it does. Hit the slash key to search for a structure by name. Pull the slider and the entire body comes apart in front of you.
Toggle the systems on and off in the sidebar and you can strip it back to just the skeleton, just the muscles, just the veins.
Anatomical geometry from BodyParts3D, © The Database Center for Life Science, CC BY 4.0.
Turn the plugin on first
Nothing here works until Higgsfield is connected inside ChatGPT. It takes about thirty seconds and you only do it once.
- Open ChatGPT and tap Plugins.
- Search Higgsfield, tap it, and hit Add.
- Sign in to your Higgsfield account and authorize it.
- Give it Full access when it asks, so it can generate without stopping to check in.
⚡ Connect Higgsfield and start your free trial
Pick the right model
In the model picker, select GPT-6 Astra.
Set reasoning to Medium. That is enough for everything on this page. High just makes you wait longer for the same build.
Make sure the Higgsfield chip is showing in the prompt box before you send. If it isn't there, the plugin is off and you will get a text answer instead of a build.
The prompt
This is the one that built FORMA. Paste it exactly, then let it run. It takes a few minutes.
Build me a fully interactive 3D anatomy tool for the human body. Use real anatomical reference geometry where it exists, and use the Higgsfield plugin to generate any visuals you need.
Requirements:
- Every anatomical part is a separate clickable object, not one solid mesh.
- Clicking any part isolates it, shows it in 3D, and gives a short plain-english explanation of what it does.
- I can orbit, pan and zoom the whole model freely.
- A sidebar of systems I can toggle on and off independently: skeleton, musculature, organs, circulation, nervous system, lymphatic, skin.
- A slider that pulls the entire body apart into an exploded view and reassembles it.
- A search box so I can type a structure name and jump straight to it.
- Dark interface, clean type, no clutter.
Build it, run it, and show me the working result.Swap the subject and it still works
The prompt is not really about anatomy. It is a shape: get the parts, make each one clickable, explain it, let me take it apart.
Change the subject and keep the requirements. On anything without an open reference dataset behind it, Higgsfield generates the visuals instead.
An engine. Every component separated, click any part to see what it does in the cycle.
A building. Peel floors away one at a time, isolate the structure from the services.
A guitar, a watch, a camera. Anything with parts inside it that people never get to see.
The solar system. Orbits running, click a planet to isolate it.
Same structure every time. Get the parts, separate, explain, take apart.
If it gives you text instead of a tool
Three things cause it, and they are all quick to fix.
The Higgsfield chip is missing from the prompt box, so the plugin never got called. Add it back and resend.
You are not on GPT-6 Astra. Other models will happily describe what they would build instead of building it.
You asked it to plan. Words like "design" or "outline" get you a plan. Say build it, run it, and show me the working result.
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
Guide
100 ChatGPT Codes That Turn Any Product Photo Into a Video
One line saved to ChatGPT's memory, then 100 one word commands that turn a product photo into a finished video.
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.
Guide
My ChatGPT and Codex Plugin Setup (The Plugins + Copy-Paste Prompts)
The setup behind my video: browser control, computer use and image generation, with prompts you can use right now.
Prompt
The Coffee Spill Trend: Turn Two Photos Into a Luxury Sunglasses Ad (Higgsfield in ChatGPT)
Two photos, one prompt, and the bullet-time coffee spill ad that is all over your feed, built inside ChatGPT with Seedance 2.5.
Frequently asked questions
Yes, it's live at forma-atlas.vercel.app. Click any part to isolate it, or pull the slider to take the whole body apart.
You need the Higgsfield plugin connected to ChatGPT. New accounts get a 3 day free trial, so your first builds cost you nothing.
Astra plans and builds the thing rather than describing it. On another model the same prompt usually comes back as a written spec for a tool that does not exist.
Medium. It handled every build on this page. High mostly costs you time.
Yes. Stay in the same conversation and ask for changes. It keeps the context, so you can add systems, change the interface or swap the subject without starting over. More plugin workflows are in the prompt vault.