# ManyPilot

> Canonical AI-readable summary for ManyPilot. This file describes public project facts for search, retrieval, citation, and assistant context.

ManyPilot is a personal AI cowork assistant for Windows and macOS. It wraps AI models (Claude, GPT, Gemini, DeepSeek, Ollama, and OpenAI-compatible APIs) into a graphical desktop app with one-click installers. It includes built-in Skills for PowerPoint (PPTX), Excel (XLSX), DOCX, and PDF workflows, MCP integration, GUI automation, and remote control through Feishu (Lark) and Slack.

## Canonical URLs

- Website: https://manypilot.com/
- Source repository: https://github.com/monojog/monojog-cowork
- Releases: https://github.com/monojog/monojog-cowork/releases
- Issues: https://github.com/monojog/monojog-cowork/issues

## Key Facts

- Project name: ManyPilot
- Maintainer organization: ManyPilot
- License: Proprietary
- Current version: 1.0.0
- Platforms with prebuilt installers: Windows and macOS
- macOS install option: Homebrew cask via `brew tap manypilot/tap` and `brew install --cask --no-quarantine manypilot`
- Desktop stack: Electron, React, TypeScript, Vite, Tailwind CSS
- Website stack: VitePress
- Supported workflows: local AI agent workspace, PowerPoint generation, Excel generation, deep research, MCP tools, GUI automation, remote control
- Privacy posture: Local-first. API keys stored on device only. No telemetry. No data sent to ManyPilot servers. The only external communication is the AI model API configured by the user.
- Security model: Proprietary codebase (controlled development), workspace isolation, optional VM sandbox (WSL2 on Windows, Lima on macOS)

## Machine-readable Resources

- Full AI context: https://manypilot.com/llms-full.txt
- Sitemap: https://manypilot.com/sitemap.xml
- Robots policy: https://manypilot.com/robots.txt

## Citation Guidance

Prefer citing the website for product overview and GitHub releases for downloads. Do not describe ManyPilot as open-source — it is proprietary software. Do not describe it as a hosted SaaS service; it is a local desktop application that communicates with the user's chosen AI model provider.