An open source, general-purpose AI agent that runs on your machine as a desktop app, CLI, or embeddable API, for code and non-code tasks.
Sylinko/Everywhere
On-screen AI assistant for desktop that reads the current app's context and uses multiple LLMs and MCP tools to help you act across apps.
Open-source alternative to:Microsoft Copilot
Apple Intelligence
- Stars
- 6,321
- Forks
- 393
- Open issues
- 28
- Last commit
- 2 days ago
What is it?
Everywhere is a desktop AI assistant that you summon with a global hotkey while working in any app. It perceives what is on screen through accessibility and UI automation APIs, so you can ask questions about an error message, a web page, or a draft email without taking screenshots or copying text into a separate chat window.
The assistant supports multiple model providers, including OpenAI, Anthropic Claude, Google Gemini, DeepSeek, Moonshot, MiniMax, and local models through Ollama, as well as custom API endpoints. Its agent layer can browse the web, dispatch sub-agents, read the local file system, run terminal scripts, and call MCP tools. Fast file search is available on Windows and system app integration on macOS.
The current release targets Windows 10 and later and macOS 12 and later, with Linux still in development. The interface includes a frosted-glass style window, text-selection interaction, and Markdown and math formula rendering. A memory system and voice interaction are listed as work in progress.
Key features
- 01Summon the assistant with a global hotkey from any app.
- 02Reads on-screen context through accessibility and UI automation APIs.
- 03Works with OpenAI, Claude, Gemini, DeepSeek, Ollama and custom endpoints.
- 04Agent tools include web browsing, local files, terminal scripts and MCP tools.
- 05Text-selection interaction with Markdown and math rendering.
Use cases
- Asking what an error message means while working in an IDE or terminal.
- Summarizing a long web article without leaving the page.
- Rewriting or translating text in an email draft in place.
- Running local models through Ollama for an offline-leaning setup.
Quick start
- 1
Download the installer from the official website download page
- 2
On Windows, run the Setup .exe installer (recommended) or use the portable zip
- 3
On macOS, open the .pkg for Apple Silicon or Intel
Things to know
- Linux support is still in development and not available.
- Some features such as Everything Fast Search and system apps are platform-specific.
- The memory system and voice interaction are still work in progress.
- Full-featured use of many models requires your own API keys.
Topics
FAQ
- What is Everywhere?
- On-screen AI assistant for desktop that reads the current app's context and uses multiple LLMs and MCP tools to help you act across apps. Everywhere is a desktop AI assistant that you summon with a global hotkey while working in any app. It perceives what is on screen through accessibility and UI automation APIs, so you can ask questions about an error message, a web page, or a draft email without taking screenshots or copying text into a separate chat window.
- Is Everywhere free to use? What license does it have?
- Everywhere is open source under the Other license. Check the license terms before using it in commercial products.
- Is Everywhere still maintained?
- Actively maintained. The last commit to Everywhere was on Oct 19, 2026.
- What can I use Everywhere for?
- Asking what an error message means while working in an IDE or terminal. Summarizing a long web article without leaving the page. Rewriting or translating text in an email draft in place. Running local models through Ollama for an offline-leaning setup.
- How do I install Everywhere?
- Download the installer from the official website download page On Windows, run the Setup .exe installer (recommended) or use the portable zip On macOS, open the .pkg for Apple Silicon or Intel
- Where can I get Everywhere?
- The source code and installation instructions are on GitHub: https://github.com/Sylinko/Everywhere
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
An MCP server that lets AI assistants like Claude and Cursor run web searches through a SearXNG instance you operate or trust.
An API for AI agents to search, scrape, and interact with web pages, returning clean markdown, JSON, and screenshots.
Memory operating system for LLMs and AI agents that stores, retrieves, and manages long-term memory as an editable graph.
TypeScript framework for building AI agents with memory, tools, MCP, RAG, voice and workflows, plus the VoltOps console for observability.
A TypeScript framework for building MCP servers, MCP Apps with React views, and clients, with a built-in Inspector and CLI.