SCrawler: the scope line is a refusal, and it is written in bold

The answer to adding a site is a refusal, and it is the section heading

The documentation has a heading that asks how to request a new site, and the answer under it is a single bold sentence saying the author is not currently accepting requests to develop new sites. That is the most useful thing on the page for a prospective user, because it settles the question of whether the site list will grow. What exists instead is a plugins document covering other sites, and a solution in which the supported sites that needed their own assembly are separate projects. So the growth path is closed by choice rather than by capacity, and the person who wrote it has decided that maintaining the existing list is the work. It is worth noting that the sentence sits where a contribution invitation would normally go, which tells you the project's position plainly.

The requirement is a specific framework version, and there is a script to check it

There is exactly one platform requirement line, and it is not a general statement about .NET. It names Windows 10 and 11, and .NET Framework 4.6.1 or higher, with a parenthetical that version 4.6.1 specifically must be installed rather than merely implied. The page then points at a PowerShell script in the repository's tools directory to check whether the installed framework version is compatible. Shipping a checker for your own runtime floor is a small courtesy most projects skip, and the reason it matters here is that the floor is an old desktop framework rather than the modern cross-platform one, so the gap between a machine that satisfies it and one that does not is wider than the version numbers suggest. The primary language is Visual Basic .NET, which also explains why the requirement is a desktop framework.

The first run takes everything, and every run after it takes only what is new

The download model is described in three sentences and it is the core of how the program behaves. First the program downloads the full profile of each user it is given. On subsequent runs it downloads only new posts. And it remembers which posts it has already downloaded, which is what makes the second run cheap. That local record is the whole state of the tool: there is no account, no server component and no sync, so the record of what you already have is a file on the machine that holds it. The consequence is that the first run against a heavy account is a long one, and that clearing the program's data to start clean is indistinguishable from losing track of what you already fetched.

A subscription mode fetches the post preview and skips the file

One mode in the settings is worth reading closely because it tells you what the author considers a light use. In subscription mode you add users and search queries as normal, and the program downloads the post preview but not the media file itself. That is a deliberate middle setting between not following an account at all and archiving everything from it: you get a feed of what was posted, in text, without the library of images and video growing on your disk. It exists alongside a fuller feed view of downloaded media and subscription posts, and alongside the scheduled automation that repeats the check every X minutes. For anyone who wants to know what an account posted without committing to storing it, that mode is the feature.

Two sites are marked partial, with the evidence being a chat thread link

The supported list runs to about twenty entries, and two of them carry an asterisk with the word partial attached. OnlyFans is listed that way, and JustForFans is listed as partial with a parenthetical naming a video issue, and both footnotes resolve to links into the project's chat server rather than to an entry in an issue tracker. That is a small but telling documentation choice. A known defect is being pointed at by a conversation rather than tracked somewhere searchable, which means the next person to hit it will find whatever the last person happened to paste into that thread. Each supported site also has its own settings page linked from the configuration section, covering the Reddit, Twitter, Bluesky, OnlyFans, Instagram, Threads, Facebook and JustForFans settings separately.

Plugins are compiled projects in one solution, not drop-in scripts

The tree is a Visual Studio solution, and the site list is reflected in its structure rather than in a plugin folder. Two sites have their own project directories named after them, a shared project holds the code they share, a provider project sits alongside them, and the main program and its updater are separate projects. A YouTube-specific pair of projects and a separate YouTube downloader project exist too, the latter being what the documentation links as a standalone application for YouTube alone. So adding a site means writing and compiling a project against the shared assembly, then contributing it as source. That is a higher bar than a scripting hook, and it is consistent with a decision to close the door on new sites, since the cost of each addition falls on the maintainer either way.

The root holds a comparison document, a screenshots directory and a rating request

Several of the repository's top-level entries are about presenting the program rather than running it: a document comparing this program with others, a directory of program screenshots, a frequently asked questions file, a document about how to support the project, a changelog and a contributing guide. The README opens its requests by asking readers to rate the program on two software directories rather than by describing what it does, and it prints a Bitcoin address for donations. The chat server link appears both as the project's declared homepage and as the recommended place to get help faster when something is wrong. None of that is a criticism, since a project asking for help through its own chat server rather than an issue tracker is a reasonable choice for its users, but it does mean the repository is doing outreach work as well as hosting code.

Versions are date-stamped, and three releases landed in three months of 2026

The three most recent tags read 2026.9.21.0, 2026.8.7.0 and 2026.5.13.0, so the version number is a year, a month, a day and a build counter. That scheme makes the release history self-describing without a changelog entry per version, and it also means the release cadence is legible from the tags alone: three releases in 2026, spaced about six weeks and then about three months apart, and the newest tag's timestamp is the same instant as the last recorded push. The documentation is bilingual in the sense that it carries a Chinese readme and two language markers in the header, but the guide itself is a wiki with one page per area, from the main window and users through download automation to the per-site settings, which is where the depth lives rather than in the readme.

Editorial conclusion

SCrawler suits someone on Windows who follows a large number of accounts and wants the archives to accumulate locally with new posts fetched on later runs, rather than starting from a fresh profile each time. Three things decide whether it fits. Check the platform before anything else, because the stated requirement is Windows 10 or 11 with .NET Framework 4.6.1 or newer and no other platform is mentioned anywhere, which rules it out for a server or a container. Read the per-site requirements pages before you rely on any particular site, because support is uneven and two of the listed sites are explicitly marked as partial with a footnote pointing at a chat thread about video handling. And settle the terms question yourself for each platform you point it at: the tool downloads from accounts using credentials you supply, which is why the project also ships a subscription mode that fetches previews without the media file, and that restraint is a reasonable signal of what the author expects people to use it for.

Frequently asked questions

Official sources

1. AAndyProgram/SCrawler on GitHub
2. License: GPL-3.0
3. Project website
4. README
5. Releases