출시 예정

소프트웨어 정보

Ithiel AI — Local-first AI desktop assistant

Ithiel AI is a local-first AI desktop assistant designed for people who value control, transparency, and performance. Run AI models on your own hardware, manage memory and chat history, and configure how Ithiel behaves within your system.

Key Features

• Local-first execution – Designed to prioritize on-device model running with user-controlled options.
• Model management – Select and load different model cores based on performance and hardware capability.
• Configurable memory – Multiple retention levels to balance usefulness and storage control.
• Profiles & organized chat history – Separate workspaces with predictable storage for conversations and attachments.
• Transparent downloads – Models are downloaded only when explicitly initiated by the user.

System Note

Performance depends on the AI model selected. Larger models may require increased GPU VRAM, system RAM, and storage capacity.

Early Access Information

Why Early Access?

Ithiel AI is entering Early Access to allow development to continue alongside community feedback. The intent is to refine usability, performance, and workflows based on real-world use.

What should I expect during Early Access?

The software is functional in its current state. Updates and refinements may occur over time as development continues. Some interface elements and workflows may change as usability and stability are evaluated.

There is no fixed roadmap or guaranteed feature timeline.

How can I provide feedback or report bugs?

The Steam Community Hub → Discussions section is the primary place for feedback, suggestions, and bug reports. Community input may help guide development priorities.

Will Ithiel AI change over time?

Ithiel AI is under active development. Improvements to usability, performance, and workflow clarity are planned, though specific features and timing are subject to change.

Does Ithiel AI run locally?

Ithiel AI is designed to prioritize local use on your own hardware. Some optional features (such as model downloads) require internet access, but downloads are user-initiated and visible.

What about performance and hardware requirements?

Performance depends heavily on the selected AI model. Larger models require more VRAM, RAM, and storage. Users can select model cores appropriate for their system capabilities.

Will the price change after Early Access?

The price during Early Access may differ from the price after Early Access concludes. Pricing adjustments, if any, are subject to change as development progresses.

AI 생성 콘텐츠 사용 공개

개발사는 게임에서 AI 생성 콘텐츠를 사용하는 방법을 다음과 같이 설명합니다.

This application generates text responses in real time using an on-device language model (LLM). All AI output is initiated by user prompts entered through the chat interface. The model runs locally on the user's hardware and does not require an internet connection in the Steam build. The application does not generate images or other media via AI. User inputs and generated content are stored locally on the user’s device and are not transmitted to external servers.