Build a Mechanic Hover Disc with GPT-6 Astra and V2Fun
Explore a hybrid 3D workflow that combines a V2Fun-generated seated mechanic with a procedural Three.js hover platform, presented as an assembled GLB model.
Interactive 3D preview · Build a Mechanic Hover Disc with GPT-6 Astra and V2FunDrag to rotate · Scroll to zoom
Take a Closer Look
-
01 / InputInput · Build a Mechanic Hover Disc with GPT-6 Astra and V2Fun
-
02 / ProcessProcess · Build a Mechanic Hover Disc with GPT-6 Astra and V2Fun
-
03 / FinalFinal · Build a Mechanic Hover Disc with GPT-6 Astra and V2Fun
Overview
This case pairs a cross-legged mechanic with a circular hover platform. The design brief combines organic detail—tousled hair, an expressive face, worn clothing, and folded limbs—with regular machinery built around concentric discs and cylindrical thrusters.
The supplied technical account assigns component planning and assembly calibration to GPT-6 Astra, with V2Fun generating complex forms and surface detail. The component sheet makes that division concrete: the textured character is identified as V2Fun-generated 3D, while the platform groups are labeled Three.js procedural. The broader GPT-6 Astra and V2Fun collaboration workflow explains this approach to combining generated geometry with agent-led construction.
Tools Used
- GPT-6 Astra: subject hierarchy analysis, component planning, and assembly calibration, as described in the technical account.
- V2Fun: generated visual references and the textured mechanic component, as identified in the reference and process sheets.
- Three.js: procedural structural links, body and seat, mechanical equipment, and controls/supports.
- GLB: the supplied final asset format.
Step 1 — Establish the Reference and Pose
The input sheet contains a complete concept reference, an isolated seated-character reference, and an inferred front, side, and rear study. All are labeled as V2Fun-generated references. The three-view study is explicitly AI-inferred, so its hidden-side details represent design interpretation rather than independently observed geometry.
The isolated character preserves the key pose: crossed legs, one hand extended toward a control grip, and the other resting near the knee. This separation provides a character-focused reference without the surrounding platform. The AI image generator for 3D references is relevant to this concept and reference preparation stage.
Step 2 — Separate the Character and Mechanical Groups
The exploded sheet presents five named groups: character, structural frame, body and seat, mechanical equipment, and controls and supports. Four short cylindrical thrusters appear together in the mechanical equipment group. The body group contains the layered circular platform and raised brown seating surface, while the control group shows a box-mounted upright lever.
The exploded arrangement uses display offsets to reveal the components. It illustrates their separation without establishing manufacturing joints or physical assembly tolerances.
Step 3 — Generate Organic Detail and Build Regular Geometry
V2Fun supplies the textured character, including the face, hair, vest, gloves, boots, and folded trousers. This is the case's documented use of image-to-3D model generation.
Three.js supplies the regular mechanical forms. The visible platform uses circular layers, a thick metallic outer body, repeated rectangular access plates, cylindrical propulsion housings, and a straight control shaft. Bronze-toned surfaces and blue-gray metal distinguish the main body from its rim, plates, and thrusters.
Step 4 — Assemble and Present the Delivered Model
The technical account describes returning the generated components to the agent for assembly calibration. In the final image, the mechanic sits on the raised seat with his extended hand positioned at the lever grip. The circular hull supports the composition, and three thruster housings are visible from the chosen view; the process sheet shows the four-piece set.
The final presentation identifies the image as a render of the delivered GLB after reloading, with no AI redraw. Compared with the concept reference, the final platform has cleaner metallic surfaces and a more pronounced curved brown seat border. It retains the central seated figure, circular silhouette, repeated access panels, and restrained ground shadow.
Workflow
Design brief → Generated concept and character references → Component planning → V2Fun character generation + Three.js mechanical construction → Assembly calibration → GLB export → Reloaded assembly render.
Prompt
English translation of the design intent:
An adult mechanic with tousled hair, wearing an old leather vest and loose work trousers, sits cross-legged at the center of a circular hovering platform. One hand grips a short joystick; the other rests on his knee. The platform consists of upper and lower concentric discs, a thick outer ring, and four evenly spaced short cylindrical thrusters. A slightly raised cushion sits on top, rectangular access covers repeat around the sides, and the underside has clear support relationships.
Use bronze and blue-gray metals. The platform hovers close to the ground with a soft shadow and no strong lighting effects. Create one complete standalone character-and-machine asset in a collectible-model style that combines realistic proportions with mild exaggeration, an expressive face, and natural clothing folds. Build the machinery from clear cylinders, discs, rectangular plates, straight rods, and a few regular curved tubes. Keep component boundaries and connections readable, preserving contact points with the seat, grip, and foot supports.
Show the entire centered subject with margins from a slightly elevated front three-quarter view. Use a white background, soft studio lighting, contact shadows, sharp component detail, and refined PBR materials. No text, numbers, logos, watermarks, complex scenery, or motion blur. Produce one finished reference image, not a collage or exploded view.
Result
The supplied output is 90-v2-final.glb, with MIME type model/gltf-binary and a file size of 40,824,208 bytes, approximately 40.82 MB.
The evidence supports a static display assembly combining a textured character with procedural machinery. The presentation notes that unseen surfaces are inferred. No case-specific generation time or token measurements are supplied. For context on the delivery format, see the AI 3D export guide covering GLB.
Frequently Asked Questions
What did V2Fun generate for the mechanic hover disc?
The reference sheet identifies V2Fun-generated concept, character, and inferred three-view references. The process sheet identifies the textured seated mechanic as the V2Fun-generated 3D component.
Which parts were built procedurally in Three.js?
The process sheet labels the structural frame, body and seat, mechanical equipment, and controls/supports as Three.js procedural groups. Visible examples include the circular hull, four cylindrical thrusters, and upright control lever.
Were the front, side, and rear references independently captured?
No. The reference sheet explicitly labels the three-view study as AI-inferred. It provides design guidance for additional viewpoints, not independent evidence of hidden surfaces.
How closely does the final assembly follow the concept?
It retains the cross-legged mechanic, hand-to-lever relationship, circular platform, bronze and blue-gray palette, and repeated access panels. The final platform has cleaner metal surfaces and a more pronounced curved seat border than the concept.
What final file is supplied for this case?
The supplied metadata lists 90-v2-final.glb at 40,824,208 bytes, with MIME type model/gltf-binary. The final presentation describes a static display model rendered after reloading the delivered GLB.
Explore More
-
GPT-6 Astra x V2Fun - From Code Generation to Model Collaboration | V2Fun
See how GPT-6 Astra and V2Fun combine agent reasoning, code generation, and AI 3D foundation models to build complex, structured 3D assets.
View page → -
V2Fun - AI Image Generator for 3D Creation
Create concept images, character references, and visual assets with V2Fun's AI image generator. Turn ideas into 3D-ready images for modeling, animation, and game design.
View page → -
V2Fun - AI 3D Model Generator
Create 3D models from text prompts or images with V2Fun's AI 3D model generator. Generate characters, props, game assets, and 3D-printable models faster.
View page → -
AI 3D Model Generator Export Guide: FBX, GLB, OBJ, or USDZ? | V2Fun
Use an AI 3D Model Generator with the right export format. Compare FBX, GLB, OBJ, and USDZ for Blender, game engines, animation, web, and AR.
View page →