Recently I tried to study the current status of the spam detection API - the so-called "research" is actually a simple Google search and Twitter question. Strangely, very few things come to mind when thinking about the spam detection API. I mean, some kind of paid or free URL endpoint that you can access it with a piece of text and any metadata it needs and it will tell you if that piece of text is spam. This seems to be something that a large number of users on the Internet can use, and companies of any size can monetize or provide it for free to showcase their smart computer systems.

Akismet: Industry giant

You might think Akismet is a WordPress thing, it does. It is an Automatic product and may be mainly used as a WordPress plugin. I'm using it on CSS-Tricks and so far it has blocked 1,989,326 spam messages.

It also has a common API. It has libraries for Dart, JavaScript, PHP, Python, Ruby, Go, etc., as well as other CMS plug-ins. So even if you use a different CMS or have a custom application, you can still use Akismet for spam detection.

Once you get the API key, you can use all the data it needs to send a POST request to the URL endpoint, if it is spam, it will return true; if not, it will return false.

To get better results over time, you can also submit content and tell it which is spam and which is benign content (benign content is the antonym of spam…good content).

Plino

Several people mentioned Plino to me.

There are a lot of praise here, such as it's free and returns JSON responses you may be used to. The buzzword of “machine learning” is also used here. I think it will become smarter as more and more people use it. However, there is no way to submit benign/spam content, so I don't know if this is true.

There are some other things that upset me. It obviously runs on Heroku, which is quite expensive when applied at scale, so it seems like it can disappear at any time without a pricing model. It feels a bit like a fun but abandoned side project. When I was writing this, my last submission was two years ago.

OOPSpam

OOPSpam looks very similar to Plino, but it has a pricing model, which is great. They announced their delay time, More than two seconds Developer Onar told me they have improved it to 1235ms . I'm not comparing it to other APIs, so I don't know if they are all so slow. To me, the two-second API call time seems a bit long, but maybe that's not a big deal because it's an asynchronous commit?

CleanTalk

CleanTalk has a clear pricing structure and seems to have a large number of customers, which is an advantage for me. However, the website looks a bit rough, which makes me a little worried.

(I'm sorry if this is a little rude, but it's just my psychological calculation. Good design is one of the cheapest investments a company can make to increase trust, so it makes me doubt that the company that ignores it.)

It seems they have a range of anti-spam solutions, though, which is interesting. For example, you can ask the API to see if the IP, email, or domain name is on the blacklist, which is a very straightforward way to block bad content, but is useful for preventing spam registrations (rather than just checking for text blocks). They also have a firewall solution, which is interesting for users who try to block spam before it hits their servers.

Email Options…

There are some APIs that look quite specialized for testing emails. That is, test your own email before sending it to make sure the email service doesn't treat it as spam. Here are some APIs I found while searching around:

• GlockApps API
• SpamAssassin

The above is the detailed content of Spam Detection APIs. For more information, please follow other related articles on the PHP Chinese website!