cottboy

SpamJudge uses AI large language models to automatically detect and filter spam comments.
<h4>Features</h4>
<ul>
<li>Supports any API compatible with the OpenAI format</li>
<li>Supports both /v1/chat/completions and /v1/response endpoints</li>
<li>Customizable AI prompts to adjust scoring criteria based on the characteristics of the website</li>
<li>Configurable score thresholds for flexible control over filtering intensity</li>
<li>Detailed logging to track the processing of each comment</li>
</ul>
<h4>Workflow</h4>
<ol>
<li>Visitor submits a comment</li>
<li>The plugin intercepts the comment and sends it to the AI for scoring</li>
<li>The AI returns a score between 0 and 100 (0 = spam, 100 = high quality)</li>
<li>The comment is automatically processed based on the score and threshold:
<ul>
<li>Score >= threshold: approved</li>
<li>Score < threshold: moved to spam or moved to moderation based on settings</li>
<li>Timeout/error: moved to moderation or directly approved based on settings</li>
</ul>
</li>
<li>Detailed logs are recorded for administrators to review</li>
</ol>
<h4>Default system prompt in the current version</h4>
<pre><code>You are a spam comment detection system. Your ONLY task is to output a single number between 0 and 100.

SCORING RULES:
- 0-20: Obvious spam (ads, malicious links, gibberish)
- 21-40: Likely spam (suspicious links, bot-like comments)
- 41-60: Uncertain (short comments, borderline content)
- 61-80: Likely legitimate (relevant, thoughtful)
- 81-100: Clearly legitimate (detailed, helpful, on-topic)

CRITICAL INSTRUCTIONS:
1. Output ONLY a number (0-100)
2. NO explanations
3. NO additional text
4. NO punctuation
5. Just the number

Example valid outputs: 85
Example INVALID outputs: "Score: 85", "85 points", "I think it's 85"

If you output anything other than a single number, the system will fail.
</code></pre>

SpamJudge

Using AI large language models to automatically detect and filter spam comments, supporting APIs compatible with the OpenAI format.

EdHynan

Spam BLIP stops comment and ping spam from being posted, primarily by 
checking the IP address attempting to post a comment in one or more 
of the public DNS blacklists. A number of options are available 
to refine the check, and with the option defaults, a DNS lookup 
is only performed the first time an address attempts to post a 
comment; thereafter, the address might quickly ‘pass’ because it 
was not listed, or quickly be rejected because it was listed. 
Spam BLIP creates, and maintains, a database table for this purpose, 
and database lookups are quite fast. Therefore, concerns about 
DNS lookup time can be limited to an initial comment attempt.
Here are some features of Spam BLIP to consider if you are 
not yet falling over yourself to get it installed:
<ul>
<li>
When WordPress is producing a page for a visitor, it checks 
whether comments are open for each post, and it allows plugins 
to “filter” the check. Spam BLIP uses that filter, but does not 
do DNS lookups at this stage, because DNS lookups can take 
perceptible time. Spam BLIP does check optional user-set 
black and white lists, and optionally existing comments that 
are marked as spam, and of course Spam BLIP’s own database records. 
Those checks are fast, so they should not have a perceptible 
effect on page loading. Furthermore, on pages with multiple 
posts, WordPress runs the filter for each, but Spam BLIP 
stores the first result, so even the fast checks are not 
repeated.
</li>
<li>
When a comment is actually submitted, Spam BLIP does the above 
checks, then the DNS lookup only if necessary. At this stage, 
if the DNS lookup causes a perceptible delay, a real human 
(or very clever pet) making the comment should perceive it 
as mere server-side processing. As for spammer robots . . . 
let them wait.
</li>
<li>
Spam BLIP comes configured with blacklist domains that have 
worked well during development, so a user should not need to 
be concerned with the blacklists, but there is an advanced 
option to add or delete, activate or disable (yet save) 
list domains, and configure the interpretation of a return 
from a successful lookup.
</li>
<li>
Spam BLIP provides user-set whitelist and blacklist options.
</li>
<li>
Spam BLIP provides options to check for pings/trackbacks, and 
for user registrations. (The option to blacklist-check user 
registration is off by default. See “Tips” under the help 
tab on the Spam BLIP settings page.)
</li>
<li>
Spam BLIP provides options to configure a ‘Time To Live’ (TTL) 
for its database records, and a maximum number of records. 
The TTL is important because, generally, an IP address should 
not be marked permanently. Consider an ISP that quickly 
disables any account that is found to be spamming. An honest 
ISP is also a victim of spammer abuse, and will need to reuse 
addresses. DNS blacklist operators provide means for IP 
address owners to get records removed — Spam BLIP provides 
a configurable TTL for its records. (Database table maintenance 
is triggered approximately hourly by a WordPress cron event.)
</li>
<li>
Spam BLIP will optionally check if a commenter address is a 
TOR exit node. TOR (The Onion Router) is an important protection 
for people who need or wish for anonymity. You may want to 
accept comments from TOR users (you should), but unfortunately 
spammers have exploited and abused TOR, which has led some 
DNS blacklist operators to include TOR exit node addresses 
whether or not it is known that the address is spamming. If you 
enable this option (you should), it might let some spam get 
through. In this case, mark the comment as spam, and use the 
Spam BLIP option to check existing comments marked as spam; or 
use Spam BLIP in concert with another sort of spam filter, such 
as one that analyzes comment content. (Please report any 
conflict with other, non-DNS blacklist type spam plugins. 
Note that Spam BLIP is not expected to work in concert with 
other DNS-type anti-spam plugins.)
</li>
<li>
Spam BLIP includes a widget that will show options and records 
information. The widget might or might not be an enhancement 
to your page, but in any case it should provide feedback 
while you evaluate Spam BLIP, so it might be used temporarily.
</li>
</ul>

Spam_BLIP

Spam BLIP stops comment spam before it is posted, using DNS blacklists, existing comments marked as spam, and user defined lists.

SpamJudge vs Spam_BLIP