30 focused guides connect research representations to asset craft, playable systems, engine delivery and responsible production.
The library follows a production sequence rather than a list of loosely related AI terms. Start with the deliverable, identify the uncertainty, and use the corresponding guide to define a test.
Use the generative 3D foundations guides when you need to distinguish meshes, neural fields, Gaussian splats, text-conditioned generation and image reconstruction. These choices affect editability, collision, lighting and runtime support.
Use the 3D asset-production guides for prompt structure, mesh cleanup, PBR materials, rigging and animation. Each guide turns visual approval into explicit delivery gates.
Use the worlds and systems and production and trust guides for level systems, mechanics, narrative behavior, benchmarking, engines, rights and provenance.
Filter changes only what is visible here; every guide has one stable canonical URL and a clear place in the hierarchy.
A field guide to models that synthesize shapes, scenes and spatial representations - And the work required before game use.
From language conditioning and image priors to explicit meshes, this guide explains the major text-to-3D routes without demo-page shorthand.
Choose a spatial representation by editability, runtime and gameplay needs - Not by screenshot quality.
A practical explanation of single-image and multi-view reconstruction, with capture advice and honest limits.
A reusable brief system that connects visual language to scale, topology, materials and engine constraints.
How to audit and repair geometry for deformation, shading, UVs, collision and real-time performance.
Build coherent base color, roughness, metalness and normal information instead of treating one attractive image as a material.
A cleanup path from attractive character mesh to stable skeleton, skin weights, facial shapes and runtime animation.
Use models for candidate generation inside a constraint, validation and playtest loop.
Rules, learned distributions and the hybrid systems that make world generation controllable.
Turn mechanic proposals into state, rules, edge cases, prototypes and measurable playtests.
Design bounded character, quest and dialogue systems with memory, safety, lore and failure states.
A vendor-neutral test plan based on accepted assets, edit time, control, export reliability and rights.
A controlled path from source files through import settings, prefabs, collision, LODs and validation scenes.
Import, material, collision, Nanite, LOD and validation decisions for generated meshes.
A practical record-keeping framework for teams navigating ownership, inputs, model terms and human authorship.
A defect-first Blender workflow for turning a fragile generated mesh into a reviewable game-asset source file.
Choose a 3D interchange format by the hierarchy, materials, animation and collaboration data that must survive the pipeline.
Normalize, import and validate AI-generated 3D assets in Godot without editing generated imports as if they were source files.
Reduce geometry, texture and delivery cost while protecting the silhouette, material response and target-device experience.
Turn stable geometry, naming, dependency and budget requirements into fast automated gates while keeping visual and gameplay approval human.
Turn art direction into a measurable grammar, reference kit and batch review process instead of asking repeated prompts to preserve style by chance.
Connect 3D training data, representations and evaluation metrics to the production behavior creators actually receive.
Build stable UVs and bake high-detail appearance onto a controlled game mesh without hiding geometry or color-space failures.
Make generated props placeable and interactive by validating real units, orientation, origins, collision shapes and gameplay contacts.
Turn text- or motion-conditioned output into editable clips by validating skeletons, contacts, root motion, retargeting and gameplay transitions.
Combine learned terrain proposals with deterministic hydrology, biome, road, navigation and streaming rules that keep worlds playable.
Design bounded game characters that turn generated language into permitted actions without surrendering canon, state or player safety.
Use validators and diverse agents to find reachability, difficulty and exploit failures while preserving human evidence about fun and readability.
Track raw generation, editable source, textures, exchange files, engine artifacts and approvals without losing the authoritative asset history.
Each guide names the output being evaluated, separates demonstrated capability from production inference, and ends with a task that a team can run. Sources establish technical or institutional facts; the checklist translates those facts into a game-development decision.
When two questions require the same evidence and lead to the same action, they belong on one page. A new route is justified only when it serves a different production decision, needs its own source set, and can provide a distinct example, validation method or comparison.
Related questions live together when they share the same decision and evidence. New pages are added only when they serve a distinct task, need their own source set, and cannot be answered cleanly inside an existing guide.
Read the complete editorial and spam-safety policy →