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.
Sirius speaks using pre-rendered audio clips. Every line of dialogue was written by hand; an open-weight text-to-speech model (Kokoro-82M, Apache 2.0) was used during development to record those written lines as audio, in the same way a voice actor would. No dialogue, text or imagery is generated by AI — the model performs a fixed script of authored lines, and the resulting audio files ship with the product. Nothing is generated while the software is running.
During the development of the application, artificial intelligence was utilized to generate supporting utility code and translate user interface localization options into different languages.
Some of PitWal's artwork and store images were made with help from AI tools and then finished by a person. PitWal's in-app commentary lines were also written with help from AI during development and ship as a fixed set. While you play, PitWal picks from those prewritten lines and fills in your own numbers from your own driving.
PitWal does not run any AI while you play and does not send your data to any AI service. If you want, PitWal lets you connect your own AI assistant, such as Claude or a coding agent local or paid, to read your own data on your own machine. That is optional, uses your own tool, and is never required to use PitWal.
Epoch Studio is a private writing and worldbuilding studio. When you ask it to, it can generate text suggestions and illustrations to help you develop your own manuscript. All generation runs locally on your own machine using bundled, permissively-licensed AI models — or optional cloud providers you connect with your own API key. Nothing is generated on its own: every suggestion is a deliberate action you take, and you review, edit, and decide what becomes part of your book. Your writing never leaves your device and is never used to train any AI model.
Elya uses generative AI to provide optional real-time character conversations, voice interactions, and other AI-assisted companion features. Users may connect supported third-party AI services or configure their own compatible providers. AI-generated content is produced dynamically based on user input and character settings.
本アプリのソースコードの大部分は、人間の指示とレビューのもとでAIコーディング支援ツール（LLM）を用いて作成されています。ストアページの文章作成にもAI支援を利用しています。 / Most of this application's source code was written using AI coding tools (LLMs) under human direction and review. AI assistance was also used in preparing store page text. ロゴおよびストア用画像の一部は画像生成AIで作成しています。 / Some artwork, including the logo and store images, was created with generative AI.
The optional cursor-buddy 3D character included in this application was pre-generated with Meshy AI under a paid commercial license. No other shipped content is AI-generated, and this application does not generate AI content during use. Built-in tools such as automatic subtitles run local machine-learning models on this device to process the user's own media only.
The collectible assets were generated with AI, also Jacky use AI as main brain to speak or interact with you, Jacky voice is a TTS model trained by us, official Jacky skins are NOT AI generated
This software uses generative AI to provide real-time text and speech translation. It recognizes on-screen text via OCR, then calls AI large language models that you configure (such as DeepSeek, GPT, Qwen, Doubao, etc.) or the built-in local translation model (Hy-MT2) to generate real-time translations, shown in the VR overlay. Translation results are for reference only, are not guaranteed to be accurate, and do not constitute professional advice.
Generative AI was used only as a writing aid: the first drafts of the character's short speech-bubble lines, the store page text and our support/privacy web pages were written with an LLM (Claude) and then reviewed and edited by our team. No AI-generated images, 3D models, animations, voices or music are used; the character model, motions and songs were created by people. Nothing is generated by AI at runtime.
Rippusion is a music generation tool: every track it produces is created by a generative AI model. The app bundles the Stable Audio 3 Small (Music) model from Stability AI, used under the Stability AI Community License, and runs it entirely on your own device. Generation happens offline on your GPU or CPU - no cloud service is involved, no account is required, and no data is sent anywhere. Music featured in our trailer was generated using Rippusion itself.
MemeWorks contains some pre-generated visual materials created with the assistance of generative AI tools during development. These materials include certain brand elements, sample images, and promotional video assets used for presentation and marketing purposes.
MemeWorks does not use generative AI during runtime, does not generate content for users in real time, and does not connect to any external AI services while the application is being used.
This software does not include any AI features or AI integration services. AI was used to create the promotional posters for the store, as the software interface alone was not suitable for poster design.
AI tools were used to help draft some text, such as this store page description and the in-app guide (all reviewed by the developer). The application itself does not ship AI-generated content: every measurement, analysis, and recommendation is computed locally with deterministic algorithms.
Generative AI tools were used during development to assist with creating and refining some pre-generated visual assets and promotional artwork. All AI-assisted content was reviewed before inclusion. The game does not generate AI content during gameplay and does not connect to external AI services.
The application includes a small set of pre-generated pixel-art example assets, including default cat character poses, icons, and promotional visuals. They were created with generative AI assistance during development and curated as static demonstration content shipped with the application.
The product does not generate AI content at runtime and does not embed, call, or operate a generative AI model or service. Its AI Material Assistant only provides editable prompt templates and links to third-party services outside the application. Players may import their own images, GIFs, audio, and Workshop content.
The pixel art in this game was created with an AI image generation tool (Pixellab).
Generated art is never used as-is: the style, proportions and palette standards are set by a human,
every sprite is reviewed by eye, and anything off-standard is rejected and redone. Grounding,
animation continuity and color consistency are likewise human-checked before adoption.
Game design, balance and specification decisions are made by a human. AI-assisted tools are also
used in the development process.
Nothing is generated by AI at runtime (pre-generated content only).
Generative AI tools were used to assist with portions of the game’s source code. All AI-assisted code was reviewed, modified, integrated and tested by the developer. The game’s models, artwork and other visual content were created or sourced separately and were not generated by AI. CoasterWorks does not generate AI content during gameplay or connect to external AI services.
AI used to assist in creating some pre-generated promotional artwork and visual assets used in snapWR and its store presentation. snapWR does not use generative AI to generate content during use.
The application includes a live generative AI assistance feature ("AI Assistant"). This feature generates real-time, text-based recommendations and diagnostics to help the user optimize their PC's performance, manage system resources, and improve game performance. Output is generated dynamically while the application is running, based on the user's own system data (CPU, GPU, RAM, disk usage) and locally gathered telemetry.
The AI does not generate images, audio, video, or other media assets. All generated output is text-only and is derived exclusively from the user's own input and locally measured system information. No user data is transmitted to any third-party generative AI service.
Guardrails: The AI operates fully offline and uses a constrained, rule-based response engine. It only processes input entered by the user within the application, cannot generate illegal or infringing content, and does not monitor or act on content outside the application. No external API, model, or content-generation service is used at runtime.
3D models used in the game were pre-generated or generated with the assistance of AI technology prior to build deployment. All models have been reviewed, integrated, and verified to ensure they comply with rights, quality standards, and platform policies.
All speech audio is generated at runtime by a local text-to-speech model (Kokoro) running on the player's PC. Memnon reads text that is already visible on the player's screen and speaks it aloud. It does not write stories, create images, or generate any other content. If the player connects their own cloud API key, those voices are synthesized by their chosen provider under their own account.
Some of the app's code was written with the assistance of AI coding tools. All artwork and store assets are human-made.
Rexclaw's companions are powered by generative AI, conversations (text and voice), along with files you create together are generated on demand by third party AI models. Some promotional artwork, including store banners, was created with AI assistance.
Generative AI was used as a creative tool to assist with parts of the artwork, code, and written content. All design choices, creative direction, edits, and final decisions were made by the human developer, with all AI-assisted work reviewed and refined before release.
Some in-game artwork, including assets shown in screenshots and store capsules, was created by modifying AI-generated imagery. All AI-assisted content was created and reviewed in advance, and no generative AI runs during gameplay.
Generative AI tools were used during development to assist with the creation and refinement of some static promotional artwork, interface text, and localization. All AI-assisted material was reviewed and edited by the developer before inclusion. The application does not generate content during use and does not connect to any external AI service.
MistWindow includes static visual assets — the application icon, task-tray icon, and the sprite sheets for the bundled pet characters — that were created with AI tools during development, and portions of its source code were written with the help of AI coding assistants. All of this content is pre-generated and ships as fixed files: MistWindow does not generate any content with AI while it is running, and it does not send user data to AI services.
The English interface text in this application was translated from the Japanese original with AI translation tools then reviewed and corrected by the developer.
This application was developed in collaboration with AI. The promotional video (PV) on the store page is an AI-assisted recreation of the actual application behavior, and AI tools were partially utilized in creating the PV and other visual assets. The application itself does not contain any functions that utilize AI.
AI tools were used to support portions of the development process. All product decisions, rigorous testing, review, and final releases remain under full human direction and oversight.
Evymind Clips uses AI to help you find the best moments in videos you choose. It can transcribe speech, translate subtitles, and suggest highlights with helpful titles and short summaries. You remain in control: every suggestion can be reviewed, adjusted, or removed before export.
The application only works from your source video. It does not create new footage, voices, music, or images. Depending on the feature, AI processing may take place on your computer or through our cloud service.
Generative AI tools were used only to assist with localization and the writing of text for the Steam store page. No content is generated while the product is running.
Generative AI tools were used during development to create the game's visual art assets, including character images and animation/video assets. All AI-assisted content is pre-generated before release and included in the game. The game does not use generative AI to create content during gameplay.
This game uses AI-assisted tools during development, limited exclusively to programming assistance, assistance with the creation of certain UI assets, and multilingual translation assistance. AI-generated output is not integrated directly into the game. All content is manually reviewed, edited, integrated, proofread, and tested by the development team before use.
AI is not used to create characters, character illustrations, character-related artwork, animations, music, sound effects, narrative, or other core creative content. The game does not generate content in real time while running, nor does it connect to external AI services to provide generative features to players.
A2GD does NOT use AI to convert images whatsoever. It is purely algorithmic software.
AI was used for the user interface designs—I wish I could have done it myself, but I have very little experience with web design.
Dashboard Engine uses AI only at runtime, on your own device, and bundles no pre-generated AI content.
On-device neural text-to-speech (installed by default): the app includes neural TTS voices (MeloTTS and Kokoro) that synthesize spoken audio from text locally on your PC. An English voice is bundled and installed by default; other languages download on demand. These voices produce speech at runtime — for the assistant's replies, spoken system alerts, and voice previews. They are used purely as speech output: a "voice profile" is a small set of tuning settings that reference a licensed base voice, and the app stores or uploads no voice recordings of any kind.
Optional AI assistant (off by default): if you turn it on, you connect it to a language model you run locally on your own machine or local network (for example Ollama or LM Studio). Your typed messages are sent to that local model to generate conversational replies in real time, which can be read aloud by the on-device TTS above. There is no cloud AI service, account, or API key — nothing leaves your PC.
All AI output is generated live from your own input and your own local models, so it is not authored or reviewed by the developer. No AI-generated content is pre-created or bundled with the app.
AI assistance was used to translate the user interface and this store page into non-English languages. The artwork, trailers, and original English text were created by the developer.
Store page capsule artwork was generated using AI tools. Application code was developed with AI assistance. All output was reviewed and tested by the developer.
Generative AI tools were used during development to assist with some pre-generated promotional and graphical artwork used for Cubing Forge, including certain Steam store, library, and marketing assets. All AI-assisted content is reviewed, curated, and finalized before publication. Cubing Forge does not use live generative AI during gameplay.
Yakutori uses on-device AI to translate Japanese text on your screen into English in real time. It reads the on-screen text with optical character recognition (PaddleOCR PP-OCRv5) and generates the English translation with a local large language model (Qwen3-4B). All AI runs entirely on your own computer — no cloud services and no internet connection are used, and nothing you translate ever leaves your device. As with any machine translation, the AI-generated output is not always perfect.
Sirius speaks using pre-rendered audio clips. Every line of dialogue was written by hand; an open-weight text-to-speech model (Kokoro-82M, Apache 2.0) was used during development to record those written lines as audio, in the same way a voice actor would. No dialogue, text or imagery is generated by AI — the model performs a fixed script of authored lines, and the resulting audio files ship with the product. Nothing is generated while the software is running.
During the development of the application, artificial intelligence was utilized to generate supporting utility code and translate user interface localization options into different languages.
Some of PitWal's artwork and store images were made with help from AI tools and then finished by a person. PitWal's in-app commentary lines were also written with help from AI during development and ship as a fixed set. While you play, PitWal picks from those prewritten lines and fills in your own numbers from your own driving.
PitWal does not run any AI while you play and does not send your data to any AI service. If you want, PitWal lets you connect your own AI assistant, such as Claude or a coding agent local or paid, to read your own data on your own machine. That is optional, uses your own tool, and is never required to use PitWal.
Epoch Studio is a private writing and worldbuilding studio. When you ask it to, it can generate text suggestions and illustrations to help you develop your own manuscript. All generation runs locally on your own machine using bundled, permissively-licensed AI models — or optional cloud providers you connect with your own API key. Nothing is generated on its own: every suggestion is a deliberate action you take, and you review, edit, and decide what becomes part of your book. Your writing never leaves your device and is never used to train any AI model.
Elya uses generative AI to provide optional real-time character conversations, voice interactions, and other AI-assisted companion features. Users may connect supported third-party AI services or configure their own compatible providers. AI-generated content is produced dynamically based on user input and character settings.
本アプリのソースコードの大部分は、人間の指示とレビューのもとでAIコーディング支援ツール（LLM）を用いて作成されています。ストアページの文章作成にもAI支援を利用しています。 / Most of this application's source code was written using AI coding tools (LLMs) under human direction and review. AI assistance was also used in preparing store page text. ロゴおよびストア用画像の一部は画像生成AIで作成しています。 / Some artwork, including the logo and store images, was created with generative AI.
The optional cursor-buddy 3D character included in this application was pre-generated with Meshy AI under a paid commercial license. No other shipped content is AI-generated, and this application does not generate AI content during use. Built-in tools such as automatic subtitles run local machine-learning models on this device to process the user's own media only.
The collectible assets were generated with AI, also Jacky use AI as main brain to speak or interact with you, Jacky voice is a TTS model trained by us, official Jacky skins are NOT AI generated
This software uses generative AI to provide real-time text and speech translation. It recognizes on-screen text via OCR, then calls AI large language models that you configure (such as DeepSeek, GPT, Qwen, Doubao, etc.) or the built-in local translation model (Hy-MT2) to generate real-time translations, shown in the VR overlay. Translation results are for reference only, are not guaranteed to be accurate, and do not constitute professional advice.
Generative AI was used only as a writing aid: the first drafts of the character's short speech-bubble lines, the store page text and our support/privacy web pages were written with an LLM (Claude) and then reviewed and edited by our team. No AI-generated images, 3D models, animations, voices or music are used; the character model, motions and songs were created by people. Nothing is generated by AI at runtime.
Rippusion is a music generation tool: every track it produces is created by a generative AI model. The app bundles the Stable Audio 3 Small (Music) model from Stability AI, used under the Stability AI Community License, and runs it entirely on your own device. Generation happens offline on your GPU or CPU - no cloud service is involved, no account is required, and no data is sent anywhere. Music featured in our trailer was generated using Rippusion itself.
MemeWorks contains some pre-generated visual materials created with the assistance of generative AI tools during development. These materials include certain brand elements, sample images, and promotional video assets used for presentation and marketing purposes.
MemeWorks does not use generative AI during runtime, does not generate content for users in real time, and does not connect to any external AI services while the application is being used.
This software does not include any AI features or AI integration services. AI was used to create the promotional posters for the store, as the software interface alone was not suitable for poster design.
AI tools were used to help draft some text, such as this store page description and the in-app guide (all reviewed by the developer). The application itself does not ship AI-generated content: every measurement, analysis, and recommendation is computed locally with deterministic algorithms.
Generative AI tools were used during development to assist with creating and refining some pre-generated visual assets and promotional artwork. All AI-assisted content was reviewed before inclusion. The game does not generate AI content during gameplay and does not connect to external AI services.
The application includes a small set of pre-generated pixel-art example assets, including default cat character poses, icons, and promotional visuals. They were created with generative AI assistance during development and curated as static demonstration content shipped with the application.
The product does not generate AI content at runtime and does not embed, call, or operate a generative AI model or service. Its AI Material Assistant only provides editable prompt templates and links to third-party services outside the application. Players may import their own images, GIFs, audio, and Workshop content.
The pixel art in this game was created with an AI image generation tool (Pixellab).
Generated art is never used as-is: the style, proportions and palette standards are set by a human,
every sprite is reviewed by eye, and anything off-standard is rejected and redone. Grounding,
animation continuity and color consistency are likewise human-checked before adoption.
Game design, balance and specification decisions are made by a human. AI-assisted tools are also
used in the development process.
Nothing is generated by AI at runtime (pre-generated content only).
Generative AI tools were used to assist with portions of the game’s source code. All AI-assisted code was reviewed, modified, integrated and tested by the developer. The game’s models, artwork and other visual content were created or sourced separately and were not generated by AI. CoasterWorks does not generate AI content during gameplay or connect to external AI services.
AI used to assist in creating some pre-generated promotional artwork and visual assets used in snapWR and its store presentation. snapWR does not use generative AI to generate content during use.
The application includes a live generative AI assistance feature ("AI Assistant"). This feature generates real-time, text-based recommendations and diagnostics to help the user optimize their PC's performance, manage system resources, and improve game performance. Output is generated dynamically while the application is running, based on the user's own system data (CPU, GPU, RAM, disk usage) and locally gathered telemetry.
The AI does not generate images, audio, video, or other media assets. All generated output is text-only and is derived exclusively from the user's own input and locally measured system information. No user data is transmitted to any third-party generative AI service.
Guardrails: The AI operates fully offline and uses a constrained, rule-based response engine. It only processes input entered by the user within the application, cannot generate illegal or infringing content, and does not monitor or act on content outside the application. No external API, model, or content-generation service is used at runtime.
3D models used in the game were pre-generated or generated with the assistance of AI technology prior to build deployment. All models have been reviewed, integrated, and verified to ensure they comply with rights, quality standards, and platform policies.
All speech audio is generated at runtime by a local text-to-speech model (Kokoro) running on the player's PC. Memnon reads text that is already visible on the player's screen and speaks it aloud. It does not write stories, create images, or generate any other content. If the player connects their own cloud API key, those voices are synthesized by their chosen provider under their own account.
Some of the app's code was written with the assistance of AI coding tools. All artwork and store assets are human-made.
Rexclaw's companions are powered by generative AI, conversations (text and voice), along with files you create together are generated on demand by third party AI models. Some promotional artwork, including store banners, was created with AI assistance.
Generative AI was used as a creative tool to assist with parts of the artwork, code, and written content. All design choices, creative direction, edits, and final decisions were made by the human developer, with all AI-assisted work reviewed and refined before release.
Some in-game artwork, including assets shown in screenshots and store capsules, was created by modifying AI-generated imagery. All AI-assisted content was created and reviewed in advance, and no generative AI runs during gameplay.
Generative AI tools were used during development to assist with the creation and refinement of some static promotional artwork, interface text, and localization. All AI-assisted material was reviewed and edited by the developer before inclusion. The application does not generate content during use and does not connect to any external AI service.
MistWindow includes static visual assets — the application icon, task-tray icon, and the sprite sheets for the bundled pet characters — that were created with AI tools during development, and portions of its source code were written with the help of AI coding assistants. All of this content is pre-generated and ships as fixed files: MistWindow does not generate any content with AI while it is running, and it does not send user data to AI services.
The English interface text in this application was translated from the Japanese original with AI translation tools then reviewed and corrected by the developer.
This application was developed in collaboration with AI. The promotional video (PV) on the store page is an AI-assisted recreation of the actual application behavior, and AI tools were partially utilized in creating the PV and other visual assets. The application itself does not contain any functions that utilize AI.
AI tools were used to support portions of the development process. All product decisions, rigorous testing, review, and final releases remain under full human direction and oversight.
Evymind Clips uses AI to help you find the best moments in videos you choose. It can transcribe speech, translate subtitles, and suggest highlights with helpful titles and short summaries. You remain in control: every suggestion can be reviewed, adjusted, or removed before export.
The application only works from your source video. It does not create new footage, voices, music, or images. Depending on the feature, AI processing may take place on your computer or through our cloud service.
Generative AI tools were used only to assist with localization and the writing of text for the Steam store page. No content is generated while the product is running.
Generative AI tools were used during development to create the game's visual art assets, including character images and animation/video assets. All AI-assisted content is pre-generated before release and included in the game. The game does not use generative AI to create content during gameplay.
This game uses AI-assisted tools during development, limited exclusively to programming assistance, assistance with the creation of certain UI assets, and multilingual translation assistance. AI-generated output is not integrated directly into the game. All content is manually reviewed, edited, integrated, proofread, and tested by the development team before use.
AI is not used to create characters, character illustrations, character-related artwork, animations, music, sound effects, narrative, or other core creative content. The game does not generate content in real time while running, nor does it connect to external AI services to provide generative features to players.
A2GD does NOT use AI to convert images whatsoever. It is purely algorithmic software.
AI was used for the user interface designs—I wish I could have done it myself, but I have very little experience with web design.
Dashboard Engine uses AI only at runtime, on your own device, and bundles no pre-generated AI content.
On-device neural text-to-speech (installed by default): the app includes neural TTS voices (MeloTTS and Kokoro) that synthesize spoken audio from text locally on your PC. An English voice is bundled and installed by default; other languages download on demand. These voices produce speech at runtime — for the assistant's replies, spoken system alerts, and voice previews. They are used purely as speech output: a "voice profile" is a small set of tuning settings that reference a licensed base voice, and the app stores or uploads no voice recordings of any kind.
Optional AI assistant (off by default): if you turn it on, you connect it to a language model you run locally on your own machine or local network (for example Ollama or LM Studio). Your typed messages are sent to that local model to generate conversational replies in real time, which can be read aloud by the on-device TTS above. There is no cloud AI service, account, or API key — nothing leaves your PC.
All AI output is generated live from your own input and your own local models, so it is not authored or reviewed by the developer. No AI-generated content is pre-created or bundled with the app.
AI assistance was used to translate the user interface and this store page into non-English languages. The artwork, trailers, and original English text were created by the developer.
Store page capsule artwork was generated using AI tools. Application code was developed with AI assistance. All output was reviewed and tested by the developer.
Generative AI tools were used during development to assist with some pre-generated promotional and graphical artwork used for Cubing Forge, including certain Steam store, library, and marketing assets. All AI-assisted content is reviewed, curated, and finalized before publication. Cubing Forge does not use live generative AI during gameplay.
Yakutori uses on-device AI to translate Japanese text on your screen into English in real time. It reads the on-screen text with optical character recognition (PaddleOCR PP-OCRv5) and generates the English translation with a local large language model (Qwen3-4B). All AI runs entirely on your own computer — no cloud services and no internet connection are used, and nothing you translate ever leaves your device. As with any machine translation, the AI-generated output is not always perfect.
