Select text in any app, AI responds in milliseconds.
Explain, summarize, rewrite, translate — all at your fingertips.
Global text selection detection across all application boundaries. Select and trigger — zero-latency floating toolbar. Powered by Windows UIA / macOS AX native APIs.
5 built-in actions + unlimited custom templates.
Your AI, your rules.
SSE streaming output with visible deep thinking chains.
Each theme crafted with Morandi color theory. Tools as art.
Alt+Number hotkeys — summon AI in an instant.
High-performance Rust backend + Tauri v2 native framework. Zero compromise.
Highlight text in any application — browsers, editors, PDFs, emails. AtHandAI detects the selection instantly and presents a floating toolbar.
Pick an AI action from the toolbar: explain, summarize, rewrite, code analysis, translate. Or run any custom prompt template you've defined.
AI streams its response word by word. Thinking process and reply are displayed separately. Deep thinking mode reveals the full reasoning chain.
Same question,
multiple models,
real-time battle.
Query 2-5 large language models simultaneously and compare response quality in real time. Let AI models review, score, and rank each other — let AI be the judge.
19 designer themes — each one a statement of aesthetics.
AtHandAI is a desktop AI productivity tool that triggers AI actions through text selection in any application. It supports explain, summarize, rewrite, translate, and more — putting AI at your fingertips.
AtHandAI supports both Windows and macOS, built with Tauri v2 and Rust for native-level performance.
AtHandAI offers free download and premium feature trials. You can upgrade at your own choice.
AtHandAI supports GPT-4o, Claude, DeepSeek, Gemini, Kimi, Qwen, GLM, Grok, MiniMax, and more. The AI Arena feature lets you compare multiple models in real time.
Simply select text in any application. AtHandAI automatically detects the selection and displays a floating toolbar. Choose an AI action to receive a streaming response, with optional deep thinking mode for full reasoning chains.
Absolutely. All data is stored locally, network transmissions are HTTPS encrypted, and text is only transmitted temporarily during AI processing — never persisted on our servers. Login credentials are encrypted via OS-level secure storage.
AI features (explain, summarize, translate, etc.) require an internet connection. Local features like excerpt management and theme switching work offline.
Yes. Beyond the 5 built-in actions, you can create unlimited custom actions with your own prompt templates to build your personalized AI workflow.