dream-num/dsh-univer-office

Give DeepSeek Harness a real office environment. Univer Office Plugin brings spreadsheets, docs, slides, canvases, relational tables, and more into one runtime — with connected data, validation, versioned changes, and isolated worktrees for multi-agent collaboration.

What it solves

This project integrates the Univer office suite into the DeepSeek Harness (DSH) environment, allowing users to create, edit, and manage office documents—including spreadsheets, text documents, presentations, and databases—directly through natural language requests to an AI agent.

How it works

The plugin acts as an interface between the DeepSeek Harness agent and the Univer office engine. When a user makes a request, the agent creates an isolated draft (worktree) to perform edits. Users can monitor the agent's progress in real-time via a live preview window, review changes within the conversation, and choose to approve or discard them. It supports importing and exporting standard Office formats like .xlsx, .docx, and .pptx.

Who it’s for

It is designed for users of the DeepSeek Harness platform who need to automate document creation, data analysis, or presentation building within a collaborative, agent-driven workflow.

Highlights

  • Supports creation and editing of Sheets, Docs, Slides, Bases, and Boards in a unified runtime.
  • Enables safe, iterative editing through isolated drafts and explicit user approval workflows.
  • Provides semantic comparison tools to inspect changes between document versions side-by-side.
  • Allows seamless import and export of common Microsoft Office file formats.
  • Features live, interactive previews of agent work directly within the chat interface.

Related

  • Project
    dream-num/univer-workspace

    Univer Workspace is an open‑source, self‑hosted office suite (Sheets, Docs, Slides, tables, boards) that integrates AI agents via a local desktop app and a CLI. Agents edit documents in isolated “Worktrees,” then humans review and merge the changes. The system consists of a Browser UI, a central Server (OpenAPI‑driven), an Agent app built on DeepSeek Harness, and a CLI for automation. It supports fine‑grained access control, custom auth, and can be run with Node 24+ and pnpm 11.

  • Project
    hust-open-atom-club/oh-dsh

    A unified runtime and workbench for AI agents that provides Desktop, Web, and TUI interfaces for managing local development tools and plugins.

  • Project
    PC2005-cloud/dsh-pet

    A desktop companion plugin for DeepSeek Harness that provides an interactive, animated pet capable of roaming the browser and desktop while reflecting AI agent status and account information.

  • Project
    Devin-AXIS/deepseek-design

    A native visual design system for DeepSeek Harness that allows AI to generate and manually refine editable websites, presentations, and videos.

  • Project
    bowenliang123/dsh-context

    dsh-context is a DeepSeek Harness plugin that adds UI panels and a slash command for deep inspection of an LLM agent’s context window, token usage, cost, and file activity, with dashboards for session‑level insights, per‑turn trends, and agent‑family (fleet) monitoring.