DeepSeek Cowork is a Windows desktop agent framework built on DeepSeek-V3.2 Interleaved Chain-of-Thought. It combines reasoning with tool use to plan, execute, and refine tasks across files, apps, and workflows in a secure desktop environment.
# DeepSeek Cowork [中文文档](README_CN.md) | [English](README.md) **DeepSeek Cowork** is a Windows desktop agent framework built on **DeepSeek-V3.2 Interleaved Chain-of-Thought**. It combines reasoning with tool use to plan, execute, and refine tasks across files, apps, and workflows in a secure desktop environment.(**This project is not an official DeepSeek development. It is a purely personal exploration driven by individual interest.**) Built by **deepseek-cowork team**.    ## 🚀 Key Features ### 🧠 Reasoning with Tool Use * **Interleaved CoT**: The agent thinks, calls tools, observes results, and continues reasoning in a single flow. * **Tool-First Exploration**: Reads real files before acting, reducing hallucinations. ### 🔌 Skill System * **Experience-First Skills**: Skills are treated as structured experience packages rather than a second execution protocol. * **Hot-Reloadable Skills**: Drop new skills into `skills/` or `ai_skills/` and use them immediately. * **Structured Experience Capture**: Runtime lessons can be stored as structured entries and synced back into `SKILL.md`. ### 🖥️ Desktop Experience * **PySide6 UI**: Modern chat bubbles, markdown rendering, and tool-call cards. * **Workspace Sidebar**: File tree and previews without leaving the app. * **Sub-Agent Monitor**: Observe parallel workers and their statuses. ### 🛰️ Daemon & IM Gateway * **Headless Daemon**: Background inference keeps UI responsive. * **Enterprise IM (Feishu)**: Send commands via IM with daily session rotation. * **Workspace Guardrails**: IM requests follow the same workspace limits unless God Mode is enabled. ## 📦 Installation ### Option 1: Run from Executable (Windows) 1. Download the latest release from [Releases](../../releases). 2. Unzip and run `deepseek-cowork.exe`. 3. No Python installation required. ### Option 2: Run from Source (Win
HAL 分层混合模型工作流 — 强模型(Claude)负责理解/拆解/验收,低成本模型(DeepSeek)负责检索/提取/清洗。Hermes Agent skill。
An LLM agent fine-tuned on DeepSeek for spaced repetition, dynamically integrating knowledge points based on the Ebbinghaus forgetting curve.
基于 STM32F103 构建的端到端 AI 智能手表生态。自研“零重定位”原生机器码动态加载引擎与页面栈式 UI 框架;集成生产级 OTA 回滚保护机制与高带宽(921600 baud)串口协议栈。通过 Node.js 中继实现 DeepSeek AI 语义控制及 ASRPRO 语音全双工交互,是一个集成了分布式计算、现代存储管理与 AI Agent 的嵌入式全栈工程。
A Meta-Agent-Driven Self-Evolving Multi-Agent System for UAV Detection and Tracking
One command to run Hermes AI Agent with a browser UI. Zero prerequisites. 一行命令,AI 就位。
网页应用Agent,接入DeepSeek、Mimo等模型