ComboX

Asking an AI about something on screen usually means a small ritual: take a screenshot, open the chatbot, paste the image, type the prompt. ComboX, presented as a Kickstarter project, is built around shortening that loop. According to the maker, it combines saved prompts with captured screen content and passes both to services such as ChatGPT, Gemini and Claude. A Chrome extension opens a supported AI chat, or reuses an existing tab, and enters the selected image and instruction.

Hardware you can steer with

The joystick is mapped to workspace navigation: the maker says it switches between Windows virtual desktops and macOS Spaces without multi-key shortcuts, and can also handle scrolling or custom actions. The rotary dial covers scrolling, media control and value adjustment, suited to tasks where turning feels more natural than tapping a key repeatedly.

Workflows, layers and integrations

• Multi-action: chain several commands into one workflow triggered by a single button
• One-tap launching of apps, folders, custom sequences and commands
• Up to 7 profiles (layers), so the same 10 keys can serve editing, AI, browsing or streaming
• Integration with supported software including Adobe Premiere Pro, Adobe After Effects and OBS Studio
• Mobile mode, extending physical shortcuts to supported smartphones and tablets, depending on each app's shortcut support

Keys, dial, joystick, profiles and actions are all configured through dedicated software, letting users arrange controls around their own habits rather than a fixed layout.

Tunable feel

The maker lists hot-swappable mechanical switches, interchangeable keycaps, customizable LCD backgrounds and adjustable key backlighting, so both typing feel and appearance can be adapted over time.