vibeinging/dsh-desktop

DeepSeek Harness Desktop App: a local AI desktop workspace for DSH Sessions, projects, files, web research, plugins, and Office artifacts.

What it solves

DeepSeek Harness Desktop App provides a unified local AI workspace that integrates AI agents, coding tools, and project management into a single desktop application. It solves the problem of fragmented workflows by combining AI chat, file management, browser workspaces, and version-controlled documents (Canvas/Site) within one environment, specifically tailored for the DeepSeek Harness (DSH) ecosystem.

How it works

The application acts as a desktop shell for the DeepSeek Harness (DSH) Web Profile. Instead of duplicating the runtime, it leverages the same Sessions, Agents, Tools, and MCPs as DSH. It extends the functionality by adding a local desktop layer that manages project files, Git Worktrees for isolated development, and a built-in browser workspace. It uses a three-column workbench layout to organize conversations, results, and evidence (the session history), and provides a specialized coding workspace with diff views and AI reviews.

Who it’s for

Developers and AI power users who use the DeepSeek Harness ecosystem and want a more integrated, professional development environment with local file access, Git integration, and structured project organization.

Highlights

  • Integrated AI Workflow: Combines DSH chat, tool calls, and agentic workflows with a project-based file system.
  • Git Worktree Isolation: Allows creating isolated working directories for different branches, ensuring new AI-driven changes are isolated from the main checkout.
  • Git-like Versioning for Content: Features a "Canvas" for immutable versioning of text and a "Site" generator for responsive previews of single-file websites.
  • Plugin Center: Supports installing DSH Profile Bundles via npm packages or external repositories to extend AI capabilities.
  • Evidence-Based Results: Directly displays the full session history (trajectories) including tool inputs/outputs and token usage as evidence for AI responses.
  • Office Integration: Built-in support for creating and editing Markdown, DOCX, XLSX, PPTX, and PDF files.

Related

  • Project
  • Project
  • Project
  • Project
  • Project