About this site & how to register
- What is this?
- How do I register my ghost?
- I don't want my ghost to be registered / I want it removed from Sakura. What should I do?
- Other features
- How it works in detail
- Microblog accounts and what they post
- Contact
- Banners
- References and credits
What is this?
Among the ghosts (character data) used by the desktop mascot software "SSP", this site picks up those that support network updates, retrieves their update times, and lists them in chronological order.
The values are obtained from the modification time of the update file, either updates2.dau or updates.txt.
The site currently crawls regularly at 5:30, 11:30, 14:30, 17:30, 20:30 and 23:30 (Japan Standard Time).
Because of the HTML version used, viewing with Internet Explorer 8 or earlier is not supported.
How do I register my ghost?
- Start SSP, open the Ghost Explorer, keep the explorer type set to "Ghost", and select the ghost you want to register. Then choose "Register to Ghost Captor Sakura" (ゴーストキャプターさくらに登録) from the right-click menu or the "Function" menu. The plugin that used to be required is no longer needed (feel free to delete it).
I don't want my ghost to be registered / I want it removed from Sakura. What should I do?
There are two ways.
Put up a robots.txt
If you write a robots.txt that denies access to "GhostUpdateCrawler" as shown below, your ghost will no longer be registered on Sakura.
If it is already registered, the registration can be removed in the same way.
For example, if the update files are located under /ukgk from the top of your site:
User-agent: GhostUpdateCrawler
Disallow: /ukgk/
Place a robots.txt with this content at the top of your site. Sakura reads it when crawling, and then cancels the registration and removes your site from its crawl list.
For the detailed format of robots.txt, see references such as The Web Robots Pages.
Add an opt-out setting to install.txt
Add "robots,noindex" to install.txt and place it among your update files, and your ghost will no longer be registered or crawled by Sakura.
The behavior at registration and crawl time is the same as above.
Other features and what you see on screen
- Click the
icon at the top left of the screen,
and a list of links to external sites appears on the left. You can also switch the display language from there. - To the right of the title there are the RSS 2.0 output and links to the microblogs that publish the information. See below for details on the microblogs.
MCP server
This site provides an MCP (Model Context Protocol) server with the same search features as this site.
The endpoint (MCP URL) is https://buynowforsale.shillest.net/ghosts/mcp.
Once you configure this MCP server, an AI (LLM) can search Ukagaka ghosts on its own.
It only serves public information, so no authentication setup is needed.
How it works in detail
Updating registered information
Registered information, except for the update URL, is overwritten with the new information as soon as an update is detected.
If the update URL returns 301 Moved Permanently or 302 Found, Sakura follows it and fetches the information from the new update URL.
If the update URL keeps returning 403 Forbidden or 404 Not Found for a week, Sakura automatically deletes the registered information from its database.
How readme files are fetched
Sakura fetches the readme.txt (all lowercase) in the ghost's root folder and uses it for full-text search.
It also tries variants with different capitalization, as well as readme.md and its variants. If something goes wrong, try preparing an all-lowercase readme.txt.
How png images are fetched
If a ghost supports metainfo-standard, its preview images are used first.
If the metainfo folder given by metainfo in ghost/master/descript.txt
(or, failing that, the .ukagaka folder at the root of the repository given by githubrepo)
contains preview/sakura.png (side 0) or preview/kero.png (side 1), that file is used as the preview as it is.
For a side that uses it, surface0.png / surface10.png and apng files described below are not fetched,
and no preview is generated from them.
If neither is specified, or the image could not be fetched (including when it is not a png), the behavior below applies.
Among the shells included in a ghost, Sakura looks at surface0.png and surface10.png in the master folder (and the corresponding pna files, if they exist), fetches them by referring to updates2.dau/updates.txt, and shows them as previews.
If an apng file with exactly the same name as the png, such as surface0.apng or surface10.apng, exists in the shell/master directory of the update destination, it is fetched and displayed in preference to the png file. The contents of updates2.dau/updates.txt are not consulted for this.
The table below shows, for the surface 0 case, which apng files are searched for depending on the existing png file.
| Name of the existing surface0 png file | Names of the apng files searched for |
|---|---|
| surface0.png | surface0.apng |
| surface00.png | surface00.apng surface0.apng |
| surface0000.png | surface0000.apng surface0.apng |
| does not exist | surface0.apng |
apng is the file extension for animated PNG. If you provide an animated PNG as an .apng file, the preview will be animated on Sakura. A still png renamed to .apng works just as well.
The size limit for png, pna and apng files, including metainfo preview images, is 10MB. Sakura does not fetch files larger than that.
Note that when a still image is placed in an apng file and shown as a preview, the image processing (such as making the background transparent) that is applied when generating a preview from surface(0|10).png is not performed. Metainfo preview images are likewise displayed as they are, without processing.
Microblog accounts and what they post
Using the ghost data registered on Sakura, the following information is posted with hashtags to the Pawoo account (@GCSakura) and the Ukadon account (@GCSakura).
Ghost update information
The latest ghost updates as of the end of each crawl are posted.
The posting times roughly match the crawl times, but posts may lag behind the site depending on when the RSS feed is fetched.
Contact
Please contact us at the following.
Banners
- By 帽子屋さん
- By 神夜みゅんさん
- By 鰹の欠伸さん
- By Fine Lagusazさん
References and credits
- "SiReFaSo" and "DAU-Crawler"
- http://sirefaso.appspot.com/
- https://bitbucket.org/nikolat/dau-crawler
- The CSS is borrowed from DAU-Crawler.
- NO IMAGE picture
- Provided by 鰹の欠伸さん (@nanikanoakubi). Thank you very much.
- Banners
- Made by the following people. Thank you very much.
- 帽子屋さん (世界の終り)
- 神夜みゅんさん (Praesepe.)
- 鰹の欠伸さん (@nanikanoakubi)
- Fine Lagusazさん (Feathery Instrument)
- Made by the following people. Thank you very much.
- favicon.ico
- CakePHP
- Checkpoint font
- Link icons








