Charisma.ai
Story-first character engine where writers keep authorial control rather than handing dialogue to a free-running model.
Visit Charisma.ai →AI tools for people who actually ship games
Game studios adopted AI for production long before it was fashionable, and the useful tools are not the ones that generate a whole game. They are the ones that remove weeks from asset pipelines, give NPCs something to say, and playtest a build a thousand times overnight. This directory covers those.
26 listings · 8 categories · updated September 2026Advertisement
Browse game AI tools
Grouped by where the tool lands in a production pipeline, from characters through assets to testing.
Convai
Conversational characters with voice and scene awareness, with plugins for the major engines.
Visit Convai →Inworld AI
Character engine for NPCs with personality, memory and goals, built to be driven from a game engine at runtime.
Visit Inworld AI →NVIDIA ACE
Suite of models for digital-human characters — speech, animation and conversation — that can run on device or in the cloud.
Visit NVIDIA ACE →Kaedim
Turns 2D concept art into production 3D models, with human review in the loop before delivery.
Visit Kaedim →Luma AI
Capture and generative 3D tooling, widely used for turning real objects and scenes into usable assets.
Visit Luma AI →Meshy
Text and image to 3D model generation with texturing, aimed at getting usable game-ready meshes quickly.
Visit Meshy →Polycam
Phone-based 3D scanning that produces meshes and splats suitable for reference or direct use.
Visit Polycam →Sloyd
Parametric 3D model generation that produces clean, low-poly, game-ready geometry rather than scan-like meshes.
Visit Sloyd →Tripo
Fast text and image to 3D generation with an API for pipeline integration.
Visit Tripo →Layer AI
Asset generation and iteration built around keeping a consistent art direction across a whole game.
Visit Layer AI →Leonardo.Ai
Image generation with fine-tuned models and controls aimed at game concept art and production assets.
Visit Leonardo.Ai →Scenario
Trains generators on your own art style so generated assets match an existing game's look instead of fighting it.
Visit Scenario →Anything World
Generates and automatically rigs animated 3D content for interactive environments.
Visit Anything World →Promethean AI
Assists artists in building virtual worlds by placing and dressing environments from a library of assets.
Visit Promethean AI →Altered
Voice transformation and performance tooling that lets one actor cover many characters.
Visit Altered →Audiokinetic Wwise
The standard interactive audio middleware — the layer any generated audio eventually has to live inside.
Visit Audiokinetic Wwise →ElevenLabs
High-quality speech synthesis and voice cloning, widely used for placeholder VO and shipped narration.
Visit ElevenLabs →Respeecher
Speech-to-speech voice cloning with a consent-based licensing model, used in film, television and games.
Visit Respeecher →Godot
Open-source engine with a community plugin ecosystem covering AI behaviour, generation and tooling.
Visit Godot →Unity
Engine with built-in neural inference for running trained models directly inside a game at runtime.
Visit Unity →Unreal Engine
Engine with machine-learning deformers and animation tooling used widely in high-end character pipelines.
Visit Unreal Engine →GameAnalytics
Free analytics built for games, with the retention and progression metrics that generic web analytics miss.
Visit GameAnalytics →modl.ai
AI agents that play your build automatically to find crashes, blockers and exploits before players do.
Visit modl.ai →Avaturn
Generates realistic rigged avatars from a photo, with SDKs for the major engines and web runtimes.
Visit Avaturn →Didimo
Generates rigged, animation-ready digital humans from a single photograph.
Visit Didimo →No matches. Try a different search.
Advertisement
Questions
Short, direct answers.
Can AI generate a whole game yet?
Not one anyone wants to play. What it genuinely does well is narrow production work — generating placeholder art, filling out barks and side dialogue, retopologising meshes, building test coverage. Treating it as an asset-pipeline accelerator produces results; treating it as a designer does not.
What are the licensing risks of AI-generated game assets?
Substantial and unsettled. Ask each vendor three things: what the model was trained on, who owns the output, and whether you are indemnified if a claim is made. Console and storefront certification may also require disclosure. Get the answers in the contract, not the marketing page, before an asset reaches a shipping build.
Are AI NPCs practical in a shipped game?
In limited roles, yes. The blockers are latency, cost per conversation at scale, and the fact that an unconstrained NPC will eventually say something that damages your rating. Studios shipping this in production constrain it hard — tight character prompts, filtered output, narrow topic scope, and a scripted fallback.
Where does AI save the most money in game production?
Quality assurance and asset iteration, not creative work. Automated playtesting explores far more of a build than a human test pass can, and generated first-pass assets let artists start from something rather than nothing. Both shorten the loop without replacing the person at the end of it.
Does this directory take payment for listings?
No. Listings are editorial and unpaid. Advertising on the site is labelled and has no bearing on what gets listed or where it appears.