Search sets for Mastodon/Fediverse

Hi,
If you are using Mastodon you are probably aware that it lacks text search. However, Google indexes the servers so it is possible to do site searches via Google. I’m wondering if it would be possible use DEVONagent to automate site search for several Mastodon servers using Search sets. I have tried but I must be doing something wrong. I have added a few servers to a Search set (with search toggles not crawl) but the set doesn’t return many results. Should I use some Google plugins with the search set? Is this a reasonable approach or not? I’m not used to building search sets so maybe I’m barking up the wrong tree :sweat_smile: Any help or comment would be appreciated.

Best regards,
Björn

1 Like

This is not a direct answer to your (great) question, but relatedly it’s easy to create an RSS feeds from Mastodon, and in particular, it can be done for hashtag searches – just append .rss to the URL. This might be useful in conjunction with DEVONagent searches.

(EDIT) Hashtag searches are relative to a server/instance, so they “see” postings from other instances that the instance is federated with. Thus, you may not have to implement searches on multiple Mastodon instances to get most postings with that tag. So for example, to search for the tag “machinelearning”, if you use a big instance like mastodon.social, you may get pretty much all postings on that topic anywhere:

https://mastodon.social/tags/machinelearning.rss

That may not be true for more obscure or specialized hashtags, in which case, you might want to add more rss feeds using different Mastodon servers/instances.

1 Like

That is a great suggestion. Leveraging rss feeds seems like a good idea. I guess one should crawl the feeds? I will try this approach also. Thanks for your reply!

Best regards,
Björn

This should be definitely possible.

Does this work as expected…
Mastodon Feeds.agentSet.zip (944 Bytes)

I have made an almost identical set as that one. Mine doesn’t follow links and have no default query. It does produce results but not so many. On the other hand I’m not sure for how long posts stay in those RSS feeds.

1 Like