Introduction

When running and debugging DeepSeek Harness (DSH) locally, developers often need to quickly see the current status: working, deep thinking, waiting for approval, task completed, or request failed. If some status prompts are placed on the web page, they would additionally rely on browser and plugin environments.

Fatfish (fatfish-pet-smart-companion) is a self-contained desktop pet: it can be used immediately after download and installation without additional configuration; if DSH is running, it can automatically switch to “real follow” mode, reflecting real-time status through expressions and bubbles.

What is Fatfish

Fatfish is a self-contained intelligent desktop pet. The repository owner is botaochen840-lgtm and the license is MIT.

It comes with all expression assets built-in, does not rely on a browser, and does not require installing any DSH plugins separately. Its main purpose is to provide a transparent, always-on-top, borderless small pet on the desktop: accompanying you quietly when idle, and following real-time status changes when running DSH.

Core Features

Status and Expressions

Supports 15 expressions/states:

  • Idle
  • Working
  • Deep Thinking
  • Waiting for Approval
  • Task Completed
  • Request Failed
  • Disappointed
  • Happy
  • Just Woke Up
  • Sleeping
  • Eating
  • Playing
  • Walking
  • Dragging
  • Welcome

A status report bubble will appear above the pet’s head when working, thinking, waiting for approval, celebrating, or encountering an error.

Desktop Behavior

  • Auto-walks when idle.
  • Randomly blinks and occasionally turns its head when stationary.
  • Automatically falls asleep and waits when idle for a long time or not connected to DSH.
  • Transparent, always-on-top, and borderless.
  • An icon appears in the bottom-right tray, with a right-click menu supporting Show, Hide, and Exit.
  • Position and size automatically restore after restart.

Interaction

  • Drag and move: You can drag the pet to other positions on the desktop.
  • Scroll wheel zoom: Supports 0.6x–2x.
  • Single-click to feed.
  • Right-click to play.

Platform and Source

  • Windows provides NSIS installer / portable version.
  • Linux provides AppImage / RPM.
  • This project is adapted from vlln’s whale-girl web plugin, and the expression assets and status semantics originate from whale-girl.
  • Whale-girl expression assets were created by ZipZipPipe.

Installation and Setup

Official executable installation commands are not provided in the documentation. You must download the installer from the release page.

# Download from the release page
https://github.com/botaochen840-lgtm/fatfish-pet-smart-companion/releases

Windows uses the NSIS installer or portable version; Linux uses AppImage or RPM. Double-click the downloaded file to run it, and the pet will appear on your desktop.

If you want to follow DSH status: After installing and running DeepSeek Harness, Fatfish will automatically detect it and switch to “real follow” mode without additional settings.

Typical Usage

  1. Download the installer for your platform from the release page.
  2. Run the installer or executable file to display the pet on the desktop.
  3. Open DeepSeek Harness when needed. Once Fatfish detects DSH, it will switch expressions and bubbles according to the real-time status.
  4. If the distance between the text and the pet becomes too large, simply drag the pet with the mouse to automatically reposition it; no restart is needed.

Use Cases and Notes

It is suitable for users who wish to keep real-time DSH status indicators on their local desktop, as well as for those who simply want to use it as a standalone desktop pet. Since it includes all expression assets built-in, it can display completely even when not connected to DSH, waiting with a sleeping animation.

In scenarios related to DSH, it should be reviewed as a local component running with the current DSH process permissions. It is recommended to check the source code, license, and origin before installation. The license is MIT, and the code was generated by AI and adapted from whale-girl.

Links

  • Community Directory Page: https://www.skillhub.cn/plugins/botaochen840-lgtm/fatfish-pet-smart-companion
  • GitHub: https://github.com/botaochen840-lgtm/fatfish-pet-smart-companion

The community directory is an independent site and has no official affiliation with DeepSeek / Synthflow. It should not be interpreted as an official app store.