Frequently Asked Questions

1. The Idea

1.1 What is Mollom?

Mollom kills unwanted content … that is, the Mollom web service helps you maintain the content quality of websites based on user-generated content. We use Machine Learning techniques, language analysis and a reputation system to ease moderation and improve the overall quality of your site’s content.

1.2 How does Mollom work (in simple terms, please)?

If a user posts a comment or other content to your site, your web server sends it to Mollom to be checked. Mollom says whether it thinks the content is spam or not and gives it a quality score. Your web server then accepts or rejects the post depending on quality standards that you set. Easy. More details can be found here.

1.3 Why care about user-generated content?

The internet has become a social meeting place, where large numbers of users share ideas, passions and information. Blogs, podcasts, forums, wikis, community sites and social networks are here to stay and attracting ever more users.

1.4 Is my site’s content quality important?

Your visitors' experiences depend entirely on the quality of your content. High quality sites attract more users and grow more quickly. Content quality also affects search engine results and, therefore, factors including the number of visitors, community growth and the amount of income that can be generated by advertising.

1.5 Will I see my site grow immediately when I use Mollom?

By using Mollom, you can allow anonymous users to post content on your site. Currently, most sites require users to create an account before they are allowed to contribute, hoping this slight barrier will improve the quality of contributed content. Enabling anonymous posting makes it much easier for people to participate in your site. So, by using Mollom you could see your site grow immediately. In the long run, the improvement in overall content quality will attract even more users to your site.

1.6 Will Mollom ever replace all my site's moderation duties?

Although Mollom is based on Machine Learning and language analysis techniques, it is not likely to replace all moderation duties a human moderator can perform. However, we can drastically reduce the effort required to moderate your site’s content by automatically separating the wheat from the chaff.

1.7 Mollom versus Akismet versus Defensio?

Mollom does offer some of the same features as Akismet or Defensio, but our goal goes further than spam-blocking alone. We want to increase the overall quality of your site's content. For example, Mollom's CAPTCHA service already helps block fake user accounts, and we are experimenting with various automated content-quality assessments, including blocking obscene, violent and profane content.

We have some great new features in the pipeline, so please check back with us regularly for more news or subscribe to our RSS feed.

2. The Product

2.1 What can Mollom currently do?

Currently, the Mollom service is an advanced spam filter, a CAPTCHA server and it provides a basic content quality score (in Beta). Many more features are on their way in the near future.

2.2 Which content management systems are currently supported?

We currently support Drupal and provide a Java library. Our open API and developer community can help you integrate Mollom into your favorite platform. Plug-ins for other content management systems, such as Wordpress, Joomla and .NetNuke are being developed and will be available soon. Contact us for details.

2.3 What’s different about Mollom’s spam filter?

Mollom uses a next generation, two-stage spam-filtering protocol. If Mollom's intelligent text analysis filter decides that some content is suspicious or if it is unsure whether the content is ham or spam, it asks the user to solve a CAPTCHA challenge before allowing the posting. This challenge/response procedure will never block human users, and it allows us to set very tight spam filters and prevent almost all spam from getting through. This mechanism all but eliminates the moderation queue. On a software level, there is also no "markHam" API call, making the service much more robust. A markHam call could allow spammers to trick users into marking spam content as ham.

2.4 Why does the spam filter return "unsure"?

If Mollom is not certain whether a post is ham or spam, it will answer "unsure," and we present the user with a CAPTCHA challenge. A human user can easily solve it and the content will be accepted, while automated spambot scripts will fail this test and be blocked. This "gray zone" allows us to use much stricter content filtering and still be sure that human users will be able to post their content.

2.5 I use Mollom and I still get spam!

Although Mollom uses the state-of-the-art in spam detection technology, it is still possible that a spam message will get through now and then. If this happens, we ask you to submit this to the Mollom service (a simple click of the "mark as abuse"). Mollom will learn from this information and instantly adjust its filters. Because all Mollom users can do this, Mollom can learn new spamming patterns very quickly. By clicking "mark as abuse", you are making sure thousands of other users will never see the spam you just marked (note that it takes more than one person to mark is as spam before Mollom considers it spam).

2.6 Where is my moderation queue?

Well … we got rid of it. We don't think you need it and here's why: The Mollom spam filter is very likely to reject spam content (currently approximately 99.8% of spam doesn't make it through) and real users will never be blocked because of the CAPTCHA challenges. Thanks to our intelligent filtering techniques, only spammers (and a very small percentage of real humans submitting suspicious content) will actually ever see the CAPTCHAs.

2.7 What is a CAPTCHA, anyway?

CAPTCHA stands for "Completely Automated Public Turing test to tell Computers and Humans Apart." It is a type of challenge-response test used to determine whether a user is human. This is done by asking a user to solve a challenge that is hard for computers, but relatively easy for human beings. Mollom currently supports both image and audio CAPTCHAs.

2.8 Aren’t CAPTCHAs an accessibility problem?

Web site accessibility is a very big issue. Adding image CAPTCHAs to web sites makes them inaccessible to visually impaired people using screen readers or braille screens. Mollom allows users to request audio CAPTCHAs, which can be solved by visually impaired users. But more importantly, only a very small percentage of users will ever see (or hear) a CAPTCHA! Using intelligent filtering techniques, we present CAPTCHAs only to users posting suspicious content. By using Mollom, your website becomes both more accessible and better-protected.

2.9 Why should I use a CAPTCHA web service and not my own home-brew ones?

CAPTCHAs are created to differentiate real human users from automated scripts. Some CAPTCHAs can be solved by advanced computer algorithms. A common workaround is the outsourcing of CAPTCHA solving to human agents in developing countries. By using a CAPTCHA web service, we can constantly monitor the quality of the CAPTCHAs we are generating and instantly adapt them if we see they have been hacked. We also use a reputation-based system to monitor the people solving CAPTCHAs, allowing us to block the outsourced human CAPTCHA solvers. Sorry for that guys ;)

2.10 Why is Mollom a web service and not a stand-alone module?

Mollom’s intelligent filter technologies learn from the spam information gathered on all sites using Mollom. This means that if another site gets spam, Mollom learns from it and protects your site better if that spam-wave rolls over you. Mollom uses so-called "crowd sourcing" or "community intelligence." The more users using Mollom, the better its performance and the faster it can learn. A stand-alone module could never achieve this.

2.11 Will Mollom be able to expand without down-time?

Mollom was built from the ground up to be extremely scalable and robust. We can instantly add servers and reroute traffic away from failed or overloaded servers to other live servers. This is made possible because the API is built on a client-side load-balancing protocol. Mollom shouldn't suffer many growing pains.

2.12 Err … What the heck is "client-side load-balancing"?

"Load balancing" means spreading around the work of running a website. High traffic websites are actually run across many servers simultaneously. "Server side" load balancing relies on a single load-balancer that distributes the work across other servers – and it can get overloaded itself. We let our clients do the load-balancing themselves by giving every Mollom client a unique list of servers that can handle its requests. When one of those servers is down or overloaded, the client can move on to the next active server on the list. This mechanism allows for very robust operations. In the future (or on request), we will even offer location-specific servers: assigning your site a preferential server near to its physical location, minimizing latency while checking the content with Mollom.

2.12 Why is the protocol based on XML-RPC?

The current API is based on XML-RPC. We chose it because it is a rich enough format to support a strong HMAC authentication scheme, and it remains relatively simple (compared to SOAP, for example). We also plan to add a REST interface in the near future and a more enterprise-grade SOAP API. Both interfaces will provide identical Mollom services while allowing you to use your favorite web service API scheme.

2.13 What about API security?

The Mollom open API uses an HMAC-based authentication scheme to make sure your key stays your key. Evil hackers cannot pluck it off the net and start checking content on your account. Since Mollom takes its enterprise users very seriously and will soon also be providing many volume-limited advanced features, it is essential that Mollom user identities are secured using an industry-grade authentication system. This is done by signing each API call to Mollom with both a private and a public key. We also add an internal reputation system on top of these keys to prevent evil users from degrading the performance of Mollom’s filters.

2.14 Do I have to use one of the clients available on your website?

No, we have a fully documented open API to access Mollom's services. You can use any of the Mollom API calls for whatever purpose may suit you. For example, if you only want CAPTCHAs, just ask for CAPTCHAs. If you only want quality checks, ignore the spam score.

2.15 Does my client have to be open source?

No, you can implement the open API and use it in your proprietary software. We do encourage openness, but you are free to use the API anywhere you like.

2.16 Can you put my home-brew client on your download page?

Please contact us and we will be happy to put your client on our download page. Please note that we will evaluate your code for correctness and general beauty. We also prefer clients that use a liberal license (GPL, BSD, etc.).

2.17 Why are most clients open source and the back-end closed source?

Mollom is built around an open API, making it possible to implement it in a large number of CMSes and web services. We strongly believe in open source and have made several open source clients available. Our back-end technology, however, will remain closed source. Giving away the intelligence we need to detect spam, generate rock-solid CAPTCHAs and figure out the quality of your content could be turned against us and defeat our purpose. The code could be used to create malicious, unblockable content.

2.18 Why doesn’t Mollom give me a precise "spaminess" or "quality" score?

Giving away the exact scores would make it easier for spammers or malicious users to fine-tune their content to get it past Mollom’s filter mechanisms.

2.19 Doesn't filtering user content jeopardize freedom of speech?

What if someone wants to have a legitimate discussion about Viagra or pornography? Those are words that can't be freely used in emails because today's email spam filters would likely block them. In other words, email anti-spam systems have effectively undermined free speech -- I cannot express myself freely without wondering if certain words are going to trigger a filter and prevent my message from being delivered to the recipients' Inboxes.

Mollom's spam blocking occurs in two steps. Only the hard-core spam (of which Mollom knows for sure that it is spam) will directly be blocked. All others posts which Mollom thinks are possibly spam, will be marked as unsure. This means that Mollom cannot be 100% sure that it is spam or not. In those cases the user will have to fill in a CAPTCHA to authenticate being a real human before the content can be accepted. More on this can be found in the How Mollom works section.

So if your users want to have a discussion about Viagra, they are free to do so, but might have to authenticate their human-ness. So Mollom fully allows free speech, between humans :)

3. The Pricing Model

3.1 How can Mollom be free, as in beer?

The basic Mollom service is free, yes, as in "free beer," without any ads or pop-ups. The free version of Mollom is limited in volume and features, but is more than adequate for the needs of your average website. This service is free because the more users we have, the more "community intelligence" we have to learn from and the more Mollom’s performance improves. Free users pay Mollom by using it and benefiting from it … great, right?

3.2 When do I have to pay for Mollom?

Do you take your site very, very seriously? Does your site generate a significant amount of advertising income? Do you need more advanced features or a higher volume of API calls? We have several subscription models to fit your needs.

3.3 What can Mollom offer enterprise users?

We strongly respect the needs of enterprise users: security, robustness, high quality of service and high volume. We can run dedicated, sole access, unlimited volume virtual servers for enterprise customers. These dedicated machines cannot be compromised by DDoS attacks since they allow only a single IP (or list of IPs) to connect to it. The machine is always 100% dedicated to your site. Using client-side load-balancing, we can even guarantee that in the unlikely event that your dedicated machine fails, you get automatically rerouted to another live Mollom server for the interim. For more information concerning dedicated Mollom servers, please contact us.

4. The Company

4.1 Mollom and Acquia?

Mollom is a self-funded, garage-style project. We take it very seriously, but it is nowhere near the size or scope of Acquia, which remains Dries's full-time commitment.

Mollom is a separate effort for two reasons: (i) it was initiated before Acquia, (ii) it is a collaboration with Benjamin Schrauwen who is not involved with Drupal or Acquia and (iii) unlike Acquia, Mollom is reaching out to as many content management systems and web applications as we can engage (and not just Drupal).

While Mollom is not associated corporately with Acquia, Acquia does intend to offer Mollom services as part of its subscription offerings.

4.2 Can I work for Mollom?

We are looking for passionate developers (both hard core back-end hacking and beautiful front-end interface designs), machine learning gurus and computational linguistics experts to join our team. We favor people who can work from our office in Antwerp, Belgium, but if you’re an absolute rock star, we might let you work remotely.

4.3 What’s with the logo?

The logo consists of a small twig growing on top of Mollom. This refers to your community web site growing and improving in peace, using the intelligent content moderation techniques provided by the Mollom web service.

4.4 How do I pronounce “Mollom”?

Mol - om

4.5 Where does the name “Mollom” come from?

It is a palindrome.
It is a Star Wars reference.
In Dutch we say "mol'em."
Mole, as in wave breaker
Mole, as in beautiful facial protrusion

Mollom