explaingit

liyupi/zhuanglema

60JavaScript
This is a quick first-pass explanation. The richer sections — use-cases, tech stack, setup, prompts — are still being generated.

TLDR

Zhuanglema (which loosely translates as "Did you install it yet?") is a desktop application that helps beginners set up software on their computers by typing a single plain-language request in Chinese.

Mindmap

A visual breakdown will appear here once this repo is fully enriched.

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

In plain English

Zhuanglema (which loosely translates as "Did you install it yet?") is a desktop application that helps beginners set up software on their computers by typing a single plain-language request in Chinese. You say something like "I want to install MySQL," and the system identifies the software, picks an appropriate version for your operating system, and generates a ready-to-run installation script with Chinese comments explaining each step. You can then run that script inside the app with one click. The project has three parts. The server is a Python FastAPI backend that handles software search, version lookup, and AI-generated installation plan storage in SQLite. The desktop application is built with Electron, React, and Vite and provides the main user interface where you search, select a version, read the generated plan, and execute it. A separate admin panel lets the project maintainer review all generated plans, handle user feedback, and trigger AI regeneration of a plan that was marked invalid. Version lookups work in tiers to avoid unnecessary AI calls: the server first checks a local cache that is valid for 24 hours, then queries official release sources for common tools like Node.js, Python, and Maven, then searches GitHub releases, and finally falls back to asking the AI model if nothing else worked. Platform detection (Windows, macOS, or Linux) is automatic but can be overridden. The generated installation plan is streamed back as Markdown using server-sent events. It includes an annotated script, an explanation of what each part does, and suggestions for what to do after installation. For software distributed through official websites rather than package managers (VS Code, IntelliJ IDEA, etc.), the plan links directly to the official download page instead of embedding a script. The app uses any OpenAI-compatible language model API, with DeepSeek-chat as the default. After installing, users can mark a plan as helpful or unhelpful and leave text comments that the admin panel aggregates for review.

Open on GitHub → Explain another repo

← liyupi on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.