# TAICHI-flet: a Windows desktop hub for wallpapers, media and AI tools > TAICHI-flet is a Flet-based Windows desktop application that bundles image, music, novel, comic and video browsing with an AI section called 论道. The source repository no longer ships the app code; the installer is what gets updated. **moshstudio/TAICHI-flet** — 基于flet的一款windows桌面应用，实现了浏览图片、音乐、小说、漫画、各种资源的功能。 - Repository: https://github.com/moshstudio/TAICHI-flet - Website: https://moshangwangluo.com - Stars: 4,737 · Forks: 500 - Language: Python - License: MIT - Published: 2026-09-23 · Updated: 2026-09-23 · Language: en - Canonical page: https://hysenlabs.com/projects/moshstudio-taichi-flet ## What TAICHI-flet bundles, and who it is built for TAICHI-flet is a Windows desktop application built on Flet. Its README describes it as a multi-purpose entertainment and creation app, and the module list backs that up: wallpapers, music search and download, web novels, comics, film and TV sources, and cloud-drive resource search. A separate section, 论道, holds the AI features: chat, a multi-model meeting room, image generation, e-commerce product images, photo editing such as ID photos and background removal, text-to-video, AI comic creation, and a Windows automation mode driven by natural language. A final module, 宝库, collects dozens of smaller tools grouped into office and media utilities, desktop widgets, AI helpers and small games. The target user is a Windows desktop user who wants these things behind one navigation bar rather than a dozen browser tabs. That is a real convenience, and it is also the main design bet: breadth over depth. If you only need one of these functions, a dedicated tool will usually do that one thing better. The value here is the aggregation and the shared account system, which the README says syncs reading records and preferences across up to three devices. ## How the Flet shell and the resource modules fit together The repository layout tells you most of what you can verify about the architecture. At the top level there is ui.py, views/, methods/, utils.py, settings.py and statics.py, with assets/ and docs/ alongside them. That is the shape of a Flet application: ui.py and views/ hold the interface, methods/ holds the logic that talks to outside services, and settings.py and statics.py hold configuration and static definitions. Flet itself renders the UI, and the pinned requirements.txt entry is flet==0.2.0.dev859, a development build rather than a stable release. The resource modules work by querying outside sources rather than hosting content. The README says 观山 lets you switch image sources, 听雨 searches multiple music sources, 观影 switches between film sources, and 搜盘 searches cloud-drive shares by keyword. Each of those implies a request layer in methods/ that hits third-party endpoints and normalizes the results for display. The AI features are different in kind: the README states they are used by points after logging in, so those calls go through an account and a metered backend rather than directly to a public API. The README does not document which backend, what the request format is, or how points are counted beyond a consumption explanation inside the app. ## Installing TAICHI-flet and taking a first pass through the modules There is no source installation path in the README. It states plainly that the open repository no longer syncs source code and that the installer keeps updating, so the documented way to get the app is the download page. The README lists the latest version as 3.5.12, dated 2026-09-07, and asks users to install the newest build for normal operation. The download link given is the project site: ```bash https://moshangwangluo.com ``` Open that page in a browser and take the PC build. The README does not list a checksum, a direct file URL, or a mirror, so the site is the only documented source. After installing, the README says a short guide appears on first launch, and that the main navigation is the fastest way to learn each module. If you want to work from the repository anyway, requirements.txt exists and pins the dependency set, including flet, requests, beautifulsoup4, lxml, opencv-python, PyMuPDF and pdf2docx. The README does not describe how to run the app from these files, and given the statement about source not being synced, treating requirements.txt as a complete build recipe would be a guess. A reasonable first session is to open 观山, switch the image source, and download one wallpaper, then open 听雨 and run one search. That exercises the request layer and confirms the app can reach its sources from your network before you invest time in the account features. ## The source-availability problem, and where the app is the wrong tool The most important limitation is stated by the project itself: the open repository no longer syncs source code. The LICENSE file is MIT, but an MIT licence on a repository that does not contain the running application's code gives you little to work with. You cannot read the current implementation, you cannot build your own binary from the published tree, and you cannot verify what the installer does beyond running it. For anyone whose adoption decision depends on code review, this rules the project out regardless of how good the app is. There are other boundaries. The app is Windows-only, so macOS and Linux users have no documented path. The AI features are metered by points and require login, and the README ties extra image, music, novel, comic, film and cloud-drive sources to a VIP subscription, listing 90+ exclusive sources. That means the free experience is narrower than the module list suggests, and the README does not publish a breakdown of which sources are free. Finally, the modules depend on third-party sites for content. The README does not describe what happens when a source changes its markup or goes offline, and there is no documented fallback beyond switching to another source where the UI offers one. ## How TAICHI-flet differs from a self-hosted media server The closest comparison is a self-hosted media server such as Jellyfin. The difference is not features, it is where the content and the logic live. Jellyfin runs on your own machine or server, indexes files you already have, and serves them to clients you control. You install it from source or a package, you own the database, and nothing in the request path depends on a third party staying online. TAICHI-flet inverts that. It ships as a closed installer, searches sources it does not host, and routes AI work through an account with points. You get a much broader catalogue without owning any of it, and you get none of the control. If your requirement is a library you curate and keep, Jellyfin is the better fit. If your requirement is to find and consume things you have not collected, and to have AI image and video tools in the same window, TAICHI-flet addresses a problem Jellyfin does not attempt. The two are not substitutes; they answer different questions about where media should live. ## Maintenance, upgrades and what the MIT licence actually covers The repository is not archived, and the last push was on 2026-09-07, so the tree is recent. That recency should not be read as active development of the application, because the README states the repository no longer syncs source. What updates is the installer, and the README asks users to stay on the latest version. In practice your upgrade path is: revisit the download page, take the new build, install over the old one. The README does not document an in-app updater, a changelog, or a rollback procedure, so if a new build misbehaves you have no documented way back to the previous one. On licensing: the repository carries MIT, which is permissive and allows reuse of what is in the tree. The README does not state the licence of the distributed installer or of the hosted AI service, and MIT on the repository does not automatically extend to those. The README also does not describe how account data or synced reading records are stored. If any of that matters to your organisation, the repository will not answer it, and this is not a question a licence file can settle. ## Conclusion TAICHI-flet suits Windows users who want one launcher for wallpapers, music, novels, comics and AI image or video tools, and who accept that the repository is a distribution point rather than a codebase. It is the wrong choice if you need to audit, patch or self-host the application, because the README states the open repository no longer syncs source. Before installing, check the version number shown on the download page against 3.5.12, and confirm which modules are usable without a VIP account. ## FAQ ### Is TAICHI-flet available for Android? The README describes TAICHI-flet as a Windows desktop application and gives a PC download, and it does not mention an Android build or an APK. The only documented distribution channel is the download page at moshangwangluo.com. ### Can I build TAICHI-flet from the GitHub repository? The README states that the open repository no longer syncs source code and that the installer keeps updating, so the published tree is not the current application. requirements.txt pins the dependency set, but the README does not describe a build or run procedure from those files. ### What is the latest version of TAICHI-flet? The README lists the latest version as 3.5.12, dated 2026-09-07, and asks users to install the newest build for normal operation. It does not provide a changelog for that release. ### Do I need a VIP account to use TAICHI-flet? The README says VIP unlocks more image, music, novel, comic, film and cloud-drive sources, described as 90+ exclusive sources, and grants extra points for the AI section. The AI features are used by points after logging in, and the README does not publish a full breakdown of which sources remain available without VIP. ## Sources - [Issues](https://github.com/moshstudio/TAICHI-flet/issues) - [License: MIT](https://github.com/moshstudio/TAICHI-flet/blob/main/LICENSE) - [moshstudio/TAICHI-flet on GitHub](https://github.com/moshstudio/TAICHI-flet) - [Project website](https://moshangwangluo.com) - [README](https://github.com/moshstudio/TAICHI-flet/blob/main/README.md) --- Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/moshstudio-taichi-flet