Public Steam disclosure research25,376 disclosures tracked
Where do games disclose AI use?
Search developers’ Steam disclosures, compare classified AI use across
games, and follow recorded changes.
This index adds cross-game classifications and statistics to
statements published on individual store pages.
Groups reflect automatic analysis of public disclosures; mature records are excluded. These counts do not measure AI use across all of Steam.
Missing or unclear statements cannot establish whether a game uses AI.
Although AI tools were used during the making of this game for efficiency, we have put our love, care, and effort into every part of it so players can experience the game as intended.
AI was used only as a support tool where helpful, and its use was kept minimal. All creative decisions, final content, and gameplay direction were reviewed and shaped by the developer to make sure nothing takes away from the intended player experience.
AI tools were used to generate the visual assets, logos, and images featured on this store page and inside the application. The software itself is a local utility and does not dynamically generate any AI content during runtime.
Pocket Chief! uses generative AI for optional text conversations, activity summaries, and task organization. Some artwork and localization used in the product and its marketing were also created with AI assistance. Users can choose Pocket Chief Cloud, connect their own compatible model provider (BYOK), or use a local model endpoint.
Portions of this game's store page text were created with the assistance of generative AI tools and then edited by the developer. CueVR itself does not use generative AI: its captions are produced entirely by on-device speech-to-text transcription of a live microphone, with no content generation and no external AI services.
Generative AI was used to assist with the creation and refinement of selected visual assets, promotional images, Steam store page graphics, and localization materials.
We want to be 100% clear: this software does not generate AI artwork. All animation, rendering, and creation tools are powered entirely by traditional algorithmic code.
The use of AI is strictly limited to:
• Text Localization: Assisted translations for store and interface text.
• Legacy Shape Detection: An optional, legacy computer-vision model that can be toggled on solely to assist with detecting edges and outlines. It does not generate art from prompts or models, and can be completely ignored in favour of our standard, non-AI outline tools.
This software uses generative AI to transcribe speech, translate live voice and text between languages, and optionally generate translated speech output. Depending on the mode selected by the user, these features may run through third-party online AI services or local models included with or configured for the software.
OverlayGG's in-game guides are written with the help of generative AI. We use large language models to research and condense publicly available information about each game into concise reference guides. All guides are generated ahead of time and stored in our catalog — nothing is generated live while you play. You can flag any guide for correction directly in the app. Some of our store-page and marketing artwork is also created using AI image tools.
This app uses our in-house AI for the pixel-art conversion process. Generative AI content is used in the trailers and screenshots shown on this store page.
Minisprite Engine was created by a small team using a combination of original work and generative AI tools.
Its mascot and hand-drawn character art were created by our designer, without the use of AI. Generative AI tools were used in creating other visual assets — including UI and store page artwork — and in drafting and localizing text across multiple languages.
All content was reviewed and refined by the team before release.
Generative AI tools, including GPT/Codex, were used during development to assist with code, UI text, layout iteration, documentation drafts, store copy, planning materials, QA checklists, and release preparation.
All AI-assisted content included in ClickAssist or its store materials was reviewed, selected, edited, and integrated by the developer before publication.
ClickAssist itself does not include live-generated AI content, AI chatbots, runtime AI generation, external AI service connections, or in-app AI image generation.
We used generative AI tools in a few clearly defined parts of this project, and we want to be upfront about where.
During development, we used AI coding assistants to help write and review parts of the source code, the same way many modern studios use these tools. We also used generative AI to help create some of the store page and marketing assets, such as promotional images and parts of the written copy. Everything produced this way was checked, edited, and approved by us before it went live.
Inside the app itself, the background removal feature runs on machine learning models to cut backgrounds out of images that you choose to load. These models only process the images you give them, on your own machine, and nothing is generated or sent anywhere without your action.
The core gameplay experience, meaning the overlays you place, the cursor effects, the pinning behavior, and the audio reactive features, is built from regular code and does not generate content with AI at runtime. Any HTML overlays you get from the Steam Workshop are made and shared by other users, so their content is up to whoever created them.
We do not use AI to generate live content that changes on its own while you use the app, and we do not train models on your data.
EchoMind uses a locally running large language model to generate conversational text responses in real time based on the user’s prompts and local memory/context.
The AI-generated content is limited to text-based conversation and local memory summaries. EchoMind does not generate images, music, voices, sound effects, 3D models, textures, or executable code as part of its normal user experience.
Generative AI was utilized to produce the mockup video game background you see on the store page. The framepacer application itself does not include AI-generated material.
FragMend is a local video-conversion utility. The application itself contains no AI-generated content and uses no generative AI at runtime, your recordings are processed entirely on your own PC by ffmpeg, with no AI involved and nothing uploaded.
Some of the text on this store page and related marketing copy was drafted with the help of a generative AI writing assistant, then reviewed and edited by the developer for accuracy. The app's logo and capsule/library artwork are generated programmatically by the developer's own code (procedural rendering), not by generative AI.
We use a local real-time deep learning neural network (AI) to perform depth estimation on the video frame. The AI analyzes the 2D image and generates a corresponding depth map in real-time, which is then used to reconstruct the 3D VR environment. The AI only processes the visual depth data and does not generate any text, audio, new art assets, or user-facing media content.
Whisper running locally on your machine is used for the speech-to-text dictation and transcribing of your own footage alone and nothing else.
No connecting to any servers, no uploading or training, and no creative decisions are handed over.
For additional languages, the UI was auto translated to bring jumpcutGG to as many people as possible. If you have any corrections, improvements, or would be interested in translating it into an additional language please contact jumpcutGG@gmail.com, thank you!
PyxelAid includes optional AI-assisted image processing filters that users can apply to their own imported images. These filters may enhance, clean, upscale, simplify, recolor, or restore pixel-art images based on the user's input. The app does not generate text, code, dialogue, or autonomous content during use. Users control the input images and decide whether to save or share the resulting outputs.
IKANDY includes an optional AI scene generator. When you choose to use it, it creates new audio-reactive visual scenes in real time from a text prompt you write. This content is generated live, only when you request it, using an AI service you connect with your own API key. Not required to use IKANDY. Some of IKANDY's artwork and promotional media (key art, trailer visuals and music, and some in-app imagery) was created with generative AI tools and is reviewed by the developer.
The store page was originally written in Chinese. All other language versions, including English, were translated using AI (the same engine the product itself uses), then reviewed by a human before publishing. No other store assets are AI-generated.
This game includes live-generated AI features that run locally on the player's computer and do not rely on external AI services.
The game uses AI to:
- interpret the player's spoken voice commands in real time to support gameplay interactions;
- generate voice audio for in-game pilot communications in real time.
The AI systems are used exclusively for these gameplay functions. They do not generate user-visible images, text, or other open-ended content.
The game is designed so that AI-generated output is limited to its intended gameplay purpose and cannot be used to create illegal, infringing, or otherwise prohibited content.
This game was developed with the extensive use of generative AI across all stages of production, including programming, visual assets, and soundscapes.
Character art and select store/marketing images were created with the help of generative AI tools, then curated and refined by our team. The application itself does not generate any content using AI at runtime.
This game uses generative AI as part of its asset production workflow. AI tools were used to help create and iterate on certain visual materials, including pixel-art characters, enemies, effects, music, and promotional images. These outputs were reviewed, selected, and integrated by the developer to match the game’s intended style and experience.
Generative AI is not used to produce personalized content from player input during gameplay, and the game does not rely on AI-generated responses or live AI systems while running.
This application uses generative AI models (Depth Anything V2 / V3) to perform real-time monocular depth estimation on user-provided videos, converting 2D footage into 3D depth maps for stereoscopic and immersive viewing. All AI models are loaded and run locally on the user's own hardware. No third-party cloud AI services are used, and no user data is transmitted externally.
Some store page and branding assets for Corner Gallery may have been created with assistance from generative AI tools. The application itself does not generate images, text, audio, code, or other AI content during normal use, and it does not connect to external AI services.
VoxelForge includes local generative AI features for user-assisted 3D and voxel asset creation.
Users can provide their own source images and use the application to generate 3D models, voxel assets, meshes, textures, and related production assets. This generation runs locally on the user’s computer using model files bundled with the application.
VoxelForge does not connect to an external AI service during normal use. It does not require a Hugging Face account, API key, cloud generation service, or external server connection at runtime. User input images and generated outputs are not uploaded to our servers.
The AI features are intended as creative tools for asset production. Users are responsible for ensuring that they have the necessary rights to any images or source material they import, and for reviewing generated outputs before using or distributing them.
RefmusePaper uses generative AI through the Refmuse online workbench to help users create and edit wallpaper images. Live AI generation requires an internet connection, a Refmuse account, and available credits. Users can enter prompts and optionally use reference images to generate or modify wallpaper artwork. Generated content is processed through the Refmuse service and is subject to safety, moderation, and abuse-prevention policies. RefmusePaper may also use AI-assisted sample artwork, screenshots, capsule/library artwork, and promotional materials to represent the wallpaper creation workflow.
CutBench is a local video tool whose core feature runs bundled, pre-trained AI models entirely on-device: they analyze the user's own locally recorded gameplay footage (visual scene understanding, speech transcription, audio event detection) and generate short text for each clip - titles, captions, and hashtags. All processing happens locally on the user's GPU; no third-party AI service is contacted and no user content is uploaded. Some store page artwork (capsules, hero, and library images) was created with the assistance of generative AI image tools and was manually reviewed and edited; screenshots are real captures of the application.
Huiyu AI Image is a creative image editor and generator powered by local AI models running on the user's own hardware. All content — including edited images and AI-generated images — is produced by diffusion-based models executed locally on the user's GPU. No content is uploaded to external servers, and all generation remains private to the user's device.
This game uses AI language models to generate real-time conversational text content during gameplay. Players interact with AI-powered desktop companions through text chat. The AI responses are generated either locally on the player's PC or through optional cloud AI services (such as OpenAI, Deepseek, etc.). All AI-generated content consists of text-based conversations only — no AI-generated images, audio, or video are used.
This work incorporates the use of generative artificial intelligence (AI) in the following areas:
Parts of the artwork, parts of the sound, and localization.
This game uses generative AI to produce dialogue content in real time. All character responses are generated by a large language model (LLM) based on user input and the loaded character card settings. No pre-written dialogue scripts are included in the shipped game. Character cards are created by third-party authors and may themselves contain AI-generated content.
AI-assisted artwork and promotional images have been used for certain store page and marketing materials. No AI-generated content is created, modified, or delivered to users during use of the software.
AI tools were used to assist in creating some store page descriptions and promotional graphics. The game itself does not use AI-generated content during gameplay.
Some of the artwork shown in our store screenshots, trailer, and on the app's in-game screen is placeholder media created with an AI image-generation tool (Google's Gemini image model). This includes the example movie posters, album covers, TV-series key art, and the cinematic "now playing" stills.
These images depict entirely fictional films, albums, and shows. They exist only as copyright-safe demo content so we can show the app's interface without displaying real, copyrighted media that the user does not own the rights to publish. They do not represent any real title.
Reel Room is a media player for the user's own library (Jellyfin server, local files, and network shares). The app performs no AI generation at runtime — all media a player sees in normal use is their own content. This is pre-generated content only; there is no live/runtime-generated AI.
Although AI tools were used during the making of this game for efficiency, we have put our love, care, and effort into every part of it so players can experience the game as intended.
AI was used only as a support tool where helpful, and its use was kept minimal. All creative decisions, final content, and gameplay direction were reviewed and shaped by the developer to make sure nothing takes away from the intended player experience.
AI tools were used to generate the visual assets, logos, and images featured on this store page and inside the application. The software itself is a local utility and does not dynamically generate any AI content during runtime.
Pocket Chief! uses generative AI for optional text conversations, activity summaries, and task organization. Some artwork and localization used in the product and its marketing were also created with AI assistance. Users can choose Pocket Chief Cloud, connect their own compatible model provider (BYOK), or use a local model endpoint.
Portions of this game's store page text were created with the assistance of generative AI tools and then edited by the developer. CueVR itself does not use generative AI: its captions are produced entirely by on-device speech-to-text transcription of a live microphone, with no content generation and no external AI services.
Generative AI was used to assist with the creation and refinement of selected visual assets, promotional images, Steam store page graphics, and localization materials.
We want to be 100% clear: this software does not generate AI artwork. All animation, rendering, and creation tools are powered entirely by traditional algorithmic code.
The use of AI is strictly limited to:
• Text Localization: Assisted translations for store and interface text.
• Legacy Shape Detection: An optional, legacy computer-vision model that can be toggled on solely to assist with detecting edges and outlines. It does not generate art from prompts or models, and can be completely ignored in favour of our standard, non-AI outline tools.
This software uses generative AI to transcribe speech, translate live voice and text between languages, and optionally generate translated speech output. Depending on the mode selected by the user, these features may run through third-party online AI services or local models included with or configured for the software.
OverlayGG's in-game guides are written with the help of generative AI. We use large language models to research and condense publicly available information about each game into concise reference guides. All guides are generated ahead of time and stored in our catalog — nothing is generated live while you play. You can flag any guide for correction directly in the app. Some of our store-page and marketing artwork is also created using AI image tools.
This app uses our in-house AI for the pixel-art conversion process. Generative AI content is used in the trailers and screenshots shown on this store page.
Minisprite Engine was created by a small team using a combination of original work and generative AI tools.
Its mascot and hand-drawn character art were created by our designer, without the use of AI. Generative AI tools were used in creating other visual assets — including UI and store page artwork — and in drafting and localizing text across multiple languages.
All content was reviewed and refined by the team before release.
Generative AI tools, including GPT/Codex, were used during development to assist with code, UI text, layout iteration, documentation drafts, store copy, planning materials, QA checklists, and release preparation.
All AI-assisted content included in ClickAssist or its store materials was reviewed, selected, edited, and integrated by the developer before publication.
ClickAssist itself does not include live-generated AI content, AI chatbots, runtime AI generation, external AI service connections, or in-app AI image generation.
We used generative AI tools in a few clearly defined parts of this project, and we want to be upfront about where.
During development, we used AI coding assistants to help write and review parts of the source code, the same way many modern studios use these tools. We also used generative AI to help create some of the store page and marketing assets, such as promotional images and parts of the written copy. Everything produced this way was checked, edited, and approved by us before it went live.
Inside the app itself, the background removal feature runs on machine learning models to cut backgrounds out of images that you choose to load. These models only process the images you give them, on your own machine, and nothing is generated or sent anywhere without your action.
The core gameplay experience, meaning the overlays you place, the cursor effects, the pinning behavior, and the audio reactive features, is built from regular code and does not generate content with AI at runtime. Any HTML overlays you get from the Steam Workshop are made and shared by other users, so their content is up to whoever created them.
We do not use AI to generate live content that changes on its own while you use the app, and we do not train models on your data.
EchoMind uses a locally running large language model to generate conversational text responses in real time based on the user’s prompts and local memory/context.
The AI-generated content is limited to text-based conversation and local memory summaries. EchoMind does not generate images, music, voices, sound effects, 3D models, textures, or executable code as part of its normal user experience.
Generative AI was utilized to produce the mockup video game background you see on the store page. The framepacer application itself does not include AI-generated material.
FragMend is a local video-conversion utility. The application itself contains no AI-generated content and uses no generative AI at runtime, your recordings are processed entirely on your own PC by ffmpeg, with no AI involved and nothing uploaded.
Some of the text on this store page and related marketing copy was drafted with the help of a generative AI writing assistant, then reviewed and edited by the developer for accuracy. The app's logo and capsule/library artwork are generated programmatically by the developer's own code (procedural rendering), not by generative AI.
We use a local real-time deep learning neural network (AI) to perform depth estimation on the video frame. The AI analyzes the 2D image and generates a corresponding depth map in real-time, which is then used to reconstruct the 3D VR environment. The AI only processes the visual depth data and does not generate any text, audio, new art assets, or user-facing media content.
Whisper running locally on your machine is used for the speech-to-text dictation and transcribing of your own footage alone and nothing else.
No connecting to any servers, no uploading or training, and no creative decisions are handed over.
For additional languages, the UI was auto translated to bring jumpcutGG to as many people as possible. If you have any corrections, improvements, or would be interested in translating it into an additional language please contact jumpcutGG@gmail.com, thank you!
PyxelAid includes optional AI-assisted image processing filters that users can apply to their own imported images. These filters may enhance, clean, upscale, simplify, recolor, or restore pixel-art images based on the user's input. The app does not generate text, code, dialogue, or autonomous content during use. Users control the input images and decide whether to save or share the resulting outputs.
IKANDY includes an optional AI scene generator. When you choose to use it, it creates new audio-reactive visual scenes in real time from a text prompt you write. This content is generated live, only when you request it, using an AI service you connect with your own API key. Not required to use IKANDY. Some of IKANDY's artwork and promotional media (key art, trailer visuals and music, and some in-app imagery) was created with generative AI tools and is reviewed by the developer.
The store page was originally written in Chinese. All other language versions, including English, were translated using AI (the same engine the product itself uses), then reviewed by a human before publishing. No other store assets are AI-generated.
This game includes live-generated AI features that run locally on the player's computer and do not rely on external AI services.
The game uses AI to:
- interpret the player's spoken voice commands in real time to support gameplay interactions;
- generate voice audio for in-game pilot communications in real time.
The AI systems are used exclusively for these gameplay functions. They do not generate user-visible images, text, or other open-ended content.
The game is designed so that AI-generated output is limited to its intended gameplay purpose and cannot be used to create illegal, infringing, or otherwise prohibited content.
This game was developed with the extensive use of generative AI across all stages of production, including programming, visual assets, and soundscapes.
Character art and select store/marketing images were created with the help of generative AI tools, then curated and refined by our team. The application itself does not generate any content using AI at runtime.
This game uses generative AI as part of its asset production workflow. AI tools were used to help create and iterate on certain visual materials, including pixel-art characters, enemies, effects, music, and promotional images. These outputs were reviewed, selected, and integrated by the developer to match the game’s intended style and experience.
Generative AI is not used to produce personalized content from player input during gameplay, and the game does not rely on AI-generated responses or live AI systems while running.
This application uses generative AI models (Depth Anything V2 / V3) to perform real-time monocular depth estimation on user-provided videos, converting 2D footage into 3D depth maps for stereoscopic and immersive viewing. All AI models are loaded and run locally on the user's own hardware. No third-party cloud AI services are used, and no user data is transmitted externally.
Some store page and branding assets for Corner Gallery may have been created with assistance from generative AI tools. The application itself does not generate images, text, audio, code, or other AI content during normal use, and it does not connect to external AI services.
VoxelForge includes local generative AI features for user-assisted 3D and voxel asset creation.
Users can provide their own source images and use the application to generate 3D models, voxel assets, meshes, textures, and related production assets. This generation runs locally on the user’s computer using model files bundled with the application.
VoxelForge does not connect to an external AI service during normal use. It does not require a Hugging Face account, API key, cloud generation service, or external server connection at runtime. User input images and generated outputs are not uploaded to our servers.
The AI features are intended as creative tools for asset production. Users are responsible for ensuring that they have the necessary rights to any images or source material they import, and for reviewing generated outputs before using or distributing them.
RefmusePaper uses generative AI through the Refmuse online workbench to help users create and edit wallpaper images. Live AI generation requires an internet connection, a Refmuse account, and available credits. Users can enter prompts and optionally use reference images to generate or modify wallpaper artwork. Generated content is processed through the Refmuse service and is subject to safety, moderation, and abuse-prevention policies. RefmusePaper may also use AI-assisted sample artwork, screenshots, capsule/library artwork, and promotional materials to represent the wallpaper creation workflow.
CutBench is a local video tool whose core feature runs bundled, pre-trained AI models entirely on-device: they analyze the user's own locally recorded gameplay footage (visual scene understanding, speech transcription, audio event detection) and generate short text for each clip - titles, captions, and hashtags. All processing happens locally on the user's GPU; no third-party AI service is contacted and no user content is uploaded. Some store page artwork (capsules, hero, and library images) was created with the assistance of generative AI image tools and was manually reviewed and edited; screenshots are real captures of the application.
Huiyu AI Image is a creative image editor and generator powered by local AI models running on the user's own hardware. All content — including edited images and AI-generated images — is produced by diffusion-based models executed locally on the user's GPU. No content is uploaded to external servers, and all generation remains private to the user's device.
This game uses AI language models to generate real-time conversational text content during gameplay. Players interact with AI-powered desktop companions through text chat. The AI responses are generated either locally on the player's PC or through optional cloud AI services (such as OpenAI, Deepseek, etc.). All AI-generated content consists of text-based conversations only — no AI-generated images, audio, or video are used.
This work incorporates the use of generative artificial intelligence (AI) in the following areas:
Parts of the artwork, parts of the sound, and localization.
This game uses generative AI to produce dialogue content in real time. All character responses are generated by a large language model (LLM) based on user input and the loaded character card settings. No pre-written dialogue scripts are included in the shipped game. Character cards are created by third-party authors and may themselves contain AI-generated content.
AI-assisted artwork and promotional images have been used for certain store page and marketing materials. No AI-generated content is created, modified, or delivered to users during use of the software.
AI tools were used to assist in creating some store page descriptions and promotional graphics. The game itself does not use AI-generated content during gameplay.
Some of the artwork shown in our store screenshots, trailer, and on the app's in-game screen is placeholder media created with an AI image-generation tool (Google's Gemini image model). This includes the example movie posters, album covers, TV-series key art, and the cinematic "now playing" stills.
These images depict entirely fictional films, albums, and shows. They exist only as copyright-safe demo content so we can show the app's interface without displaying real, copyrighted media that the user does not own the rights to publish. They do not represent any real title.
Reel Room is a media player for the user's own library (Jellyfin server, local files, and network shares). The app performs no AI generation at runtime — all media a player sees in normal use is their own content. This is pre-generated content only; there is no live/runtime-generated AI.
