Profanity Filtering & Offensive Text Moderation. Technology You Can Count On

Is letting you down?
ProfanityBlocker is the ultimate solution for content moderation.
Our intelligent engine scans every piece of user-generated content, ensuring that no profanity slips through the cracks. Our Image Moderation scanning technology also provides double the peace of mind by removing all inappropriate text from images.

Unlike other profanity filtering solutions (many profanity filtering solution's use black and white wordlist based filtering), ProfanityBlocker uses innovative technology like Machine Learning to minimise false positives and creative content filtering exploits that may frustrate your community.
Our filter platform goes beyond and understands nuance and meaning behind words.
We recognize that every project has unique community requirements, and our system is built to meet them. Trust us to provide you with the best profanity filtering solution on the market.

What's different?

Need ? We do that and more
ProfanityBlocker is an efficient and nifty text and image moderation service with a Discord bot.
Go beyond blacklist/whitelist content filtering.

Filtering done your way

Since every project has different requirements around filtering accuracy and speed, we offer several types of filtering methods, giving you full control and flexibility.

Hybrid and InsightGuard Filter has a trainable custom model on top of the base model. You can help train the base model by reporting false positives (no filter is perfect) which will get reflected at our frequent incremental learning time – custom model too based on your own criteria (learn more about custom models here).

Filter models

Choose from many filtering models for complete flexibility.

Blacklist/Whitelist

We maintain the basic blacklist for typical bad language and you can customise what you classify as bad or good with your own blacklist/whitelist.  This filtering method is the least accurate out of all the filters, but is also the fastest.

Hybrid and InsightGuard Filter support custom training on content you add from the filter logs (request history).

Hybrid Filter

  • Binary Classification, trained on several classes of data including profanity, hate speech, sexual content, violence, racism and more.
  • Training de-duplication of datasets.
  • False positive reporting and learning capability on base model and client custom model. False positive validation before/after with mitigation techniques.
  • Incremental learning means it can train faster.

This method is a suitable middle-ground for most projects as it is more intelligent and faster than blacklist/whitelist at the expensive of no real understanding of meaning when it comes to understanding context (like InsightGuard Filter or DeepSense Filter type).

This might not matter for your use case if you need a faster auto moderating experience with slightly less accuracy than type InsightGuard Filter (and may block more seemingly bad words than InsightGuard Filter type words, as it just sees a potential bad word without understanding subtext. We train the Hybrid model with vast amount of bad language, bullying, sexualisation, hate speech, racism type content and more..

InsightGuard Filter

  • Binary Classification, trained on several classes of data including profanity, hate speech, sexual content, violence, racism and more.
  • Training de-duplication of datasets.
  • False positive reporting and learning capability on base model and client custom model. False positive validation before/after with mitigation techniques.
  • Vast knowledge of the English language, understands nuance/hidden meaning behind the words.
  • Incremental learning means it can train faster.
 
It’s designed to discern context, allowing it to differentiate when more mature language is used non-maliciously. This nuanced approach may permit certain mature expressions, depending on their context and intent, such as when they’re not used in an offensive or attacking manner.
 
This system is particularly suited for mature communities that engage in complex communication. It’s adept at identifying genuinely harmful content, which goes beyond mere swear words to include bullying, sexualization, hate speech, racism, and more.

DeepSense Filter

  • Multi-layered model: toxicity and up to 7 classes, word-level,  sentence classification capability.
  • Fine-tune your class acceptability thresholds.
  • Model bias mitigation training regime, multilingual (supporting 4 languages) capable.
  • Future capable of domain specific datasets to select additional language criteria for filtering.
 
This is ideal for mature communities that have more mature and complex communication where the real bad content isn’t just with your typical swear words, but bad language, bullying, sexualisation, hate speech, racism type content and more. 

Not sure which one to use?

We recommend starting with Hybrid as it is a good middle ground. You can always change your filter options and update/downgrade your filtering plan. if you run multiple servers, you can choose the right one for your audience.

Image classifications

Text extraction

We extract the text from images and run it through the filter that your server is set to use on the premium plan

Looking for advanced Image scanning for nudity, violence and suggestive content?

We recently decided to drop efforts for content classification around nudity, violence and suggestive content for images, as Discord offers a NSFW filter.

If you want to read more about this decision and want us to revise the idea, check out this page

More on the Discord bot

The usual, kicking, banning members from your server – peace of mind from troublesome users – what every moderation bot should have!

With pb!clear – you can specify the number of messages to delete and it’s gone!

Mass Caps detection

Mass Mention detection

Link detection

Phone number detection

Use our If This Then Profanity blocks system to create rules and perform automated actions on your behalf

More on the Chromium Extension

We recognize the presence of unsettling content across various online platforms, including social media sites and blogs.

Our goal is to foster safer, more inclusive digital spaces, particularly on platforms like Discord, Reddit and Twitter/X that may lack automated moderation, by concealing NSFW material and explicit language.

Our content filters are customizable, allowing users to adjust sensitivity levels for an optimal browsing experience. Although not all administrators or owners of Discord servers, websites, forums, or blogs may choose to implement our moderation bot, we provide a viable alternative.

Our Chromium-based browser extension, in coordination with our filtering technology, scans web pages to blur content identified as inappropriate. This approach is designed to be minimally invasive, enabling users to view surrounding content and reveal the blurred content with a simple click if desired.

The extension is available for free to all users with an account, enhancing your browsing experience both within Discord and on the broader internet.

Not sure which filter to use?

We recommend starting with InsightGuard Filter. The browser extension will send a lot of text from webpages, for a faster result it is a good middle ground. You can always change your filter options and update/downgrade your filtering plan. if you want smarter filtering or category identification make sure to choose DeepSense Filter.

Blur out questionable content

Avoid surprises, we highlight the nature of the content

Simply click on blurred out text to see anyway

Keep safe on Discord

When a server doesn't use our bot, you can use our extension

Works on Reddit and hundreds of other sites

Know what you type could be bad before sending

Privacy Disclaimer:

To ensure your privacy, you have the option to disable the plugin or utilize site exclusion settings for websites you prefer not to be scanned. We retain scanned content for no more than 30 days, allowing you the opportunity to train custom models and refine your filtering experience.

Parental controls:

Parents have the option to install our plugin on their children’s browsers for enhanced online safety. If you’re interested in a stricter setup, where blurred content remains hidden and the extension cannot be uninstalled by users, please contact us. We are currently exploring interest in this feature and would appreciate your feedback.

Other Disclaimer:

A typical browsing session can saturate the free plan within minutes, and the Pro plan should be sufficient for typical browsing, but this various on how much text, and how many separate submissions are needed to be processed (i.e larger block of text will require less hourly submissions compared with Reddit/Twitter/X posts with hundreds (or thousands) of comments and replies).

To ensure a smooth service and fair usage new plans might be required for extension usage given how many requests a typical browser session can make to our filtering service, and real-time filtering can be expensive at large volumes. Should you be subscribed to Pro or Ultima plans  at which point we decide we need new plans for extension usage only, we will try and match the old plan fairly at a grandfathered rate.

Client Testimonials

“A testimonial from a client who benefited from your product or service. Testimonials can be a highly effective way of establishing credibility and increasing your company's reputation.”
Client Name
“A testimonial from a client who benefited from your product or service. Testimonials can be a highly effective way of establishing credibility and increasing your company's reputation.”
Client Name
“A testimonial from a client who benefited from your product or service. Testimonials can be a highly effective way of establishing credibility and increasing your company's reputation.”
Client Name

Pricing

Choose the plan that’s right for you and your requirements.

Each message submission counts as a request, for fair usage and the cost of running powerful AI models we must price accordingly (real-time message scanning in bulk is really power compute hungry!).

  • Free $0

    Free plan allows 200 requests an hour and 1 concurrent requests, free of charge. Hitting the limit means messages won't be scanned until the next hour.
    Includes:
    * Ability to choose between Hybrid and InsightGuard Filter and DeepSense Filter* (for now, during beta)

  • Pro $5 p/m

    Pro includes 4,000 hourly requests and 4 concurrent requests. Ideal for a medium sized Discord server.
    Includes:
    * Ability to choose between all filter models including Hybrid, InsightGuard Filter and DeepSense Filter
    * Ability to add custom message history from your Discord Server to train to block that content
    * Advanced custom rules and filter options
    * Train your own custom blocking criteria by adding chat history to a training set it learns is bad - on top of base model
    * Text scanning in images

  • Ultima $10 p/m

    Gives you 12,500 filtering requests an hour and 10 concurrent requests. Ideal for a busy Discord server.
    Includes:
    *Ability to choose between all filter modelsi including Hybrid, InsightGuard Filter and DeepSense Filter
    * Advanced custom rules and filter options
    * Train your own custom blocking criteria by adding chat history to a training set it learns is bad - on top of base model
    * Text scanning in images

Need more of image scans or specific filtering requests per hour?

Should this become normal we will look into adding a “top-up” feature, so if you need more flexible pricing, please get in touch – we realise there’s no one glove fits all, neither should our pricing.

All paid plans

include these neat perks
  • Text Scanning in Images
  • Filter options
  • Filter/chat logs
  • Custom model training (Hybrid/InsightGuard Filter)
  • No dashboard ads

A Title to Turn the Visitor Into a Lead

This is your chance to emphasize why the visitor should contact you right now.