Hold a hotkey over any content on screen to get an instant AI explanation, translation, or summary.
Use voice input and output for a hands-free question-and-answer loop while working.
Save AI answers as styled HTML, PDF, or Markdown documents for later reference.
| gonemedia/aipointer | browser-use/browsercode | ardupilot/node-mavlink | |
|---|---|---|---|
| Stars | 98 | 97 | 96 |
| Language | TypeScript | TypeScript | TypeScript |
| Last pushed | — | — | 2025-08-26 |
| Maintenance | — | — | Quiet |
| Setup difficulty | easy | moderate | moderate |
| Complexity | 2/5 | 3/5 | 3/5 |
| Audience | general | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires your own API key from a provider such as OpenRouter, Anthropic, OpenAI, or Gemini.
AIPointer is a desktop application for Windows, macOS, and Linux that lets you ask AI questions about anything on your screen by just hovering your cursor over it. Hold a keyboard shortcut, Right-Cmd on Mac, Right-Ctrl on Windows and Linux, a translucent panel opens next to your cursor, and you type your question. The app takes a screenshot of the area around your cursor, sends it along with your question to an AI service, and displays the answer, typically in under two seconds. You bring your own API key from one of several AI providers and pay directly for what you use. No data is sent to any other servers, and there is no telemetry or usage tracking. The app works completely offline from the AIPointer service's own perspective, it just connects directly to whichever AI provider you choose. Built-in slash commands let you quickly summarize, translate, explain, write code, improve text, define words, solve math problems, or identify objects. You can also ask in plain language without any special commands. Answers can be saved as HTML, PDF, or Markdown files. Additional features include voice input, so you can speak your question instead of typing, voice output that reads the answer aloud, a hands free loop mode for voice-first interaction, mouse-wiggle activation where shaking the cursor left and right quickly opens the panel, and a child safe mode with restricted content and tools. Seven built-in action tools let the AI open URLs, copy text, launch apps, and more, each one requiring your manual approval before it runs. The app is open source under the MIT license and includes an automatic updater, light and dark themes, and a settings panel for adjusting providers, hotkeys, and voice preferences.
A desktop overlay app that answers AI questions about whatever your cursor is pointing at, using your own API key and no telemetry.
Mainly TypeScript. The stack also includes TypeScript, Electron, React.
Permissive MIT license, use freely for any purpose, including commercial use.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly general.
This repo across BitVibe Labs
Verify against the repo before relying on details.