Intelligent Filtering Engine
Automated content and profanity filtering of user-generated content is a seemingly simple undertaking that quickly reveals its true complexity under the stresses of demanding, real-world deployments. First released in 2008, CleanSpeak remains the industry standard for enterprise-scale content filtering and continues to set the market benchmarks for accuracy, performance, security and customization.
Automatically Filter Content In Real-Time From Nearly Any Source
CleanSpeak can identify and block profanity, slang, racial slurs, religious terms, hate-speak, drug and alcohol terms, pornographic images and video, personally identifiable information, and other types of undesirable content and behavior from the following sources:
- Live Chat
- Forums
- Form Data
- Text Messages
- Social Networks
- Blogs
- Ratings and Reviews
- Virtual Worlds
- Live Content Feeds
CleanSpeak integrates with nearly any application and elegantly manages multiple streams of content from the same interface. Average latency is less than 1 millisecond and throughput capacity approaches 1 billion transactions per day without any extraordinary hardware or engineering requirements.
Intelligently Block Creative Attempts To Circumvent The Filter
Let’s use the word “smurf” as an example of a new word added to your base filtering list. CleanSpeak will automatically find and defeat…
- Other Parts Of Speech: smurfy | smurfer | smurflingly
- Names, Proper Nouns & User-Names: The Smurfs | Smurf2255 | BigSmurfOnline
- Odd Capitalizations: sMUrf | SmurF
- Spaces & Separators: s m u r f | smu rf | smur f
- Added Punctuation: s.m.u.r.f | s-m.u-r.f
- “Leet” Speak: s/\/\urf | sm\/rf
- Repeated Characters: sssssmmmmmuuuuurrrrfffff
- Character Replacement: $murf | sm@rf
- Phonetic Replacement: smurph | smurphy
- Misspellings & Variations: smirf | smarf
- Embedded Words: smurfhead | smurfhole | mothersmurfer | smurftastic
CleanSpeak will intelligently recognize complex combinations of the above variations. For example: “big$mmmmm\/rpher” (embedded word, character replacement, repeated characters, leet speak, phonetic replacement, and different part of speech) is automatically recognized and blocked from the single root entry of “smurf”. Enter a word once, and CleanSpeak does the rest…
Multilingual Filtering With Intelligent Language Support
- English
- Spanish
- French
- Italian
- German
- Polish
- Dutch
- Russian (in development)
- All Others As Requested
Intelligent language support lets CleanSpeak do all the work in managing your base filtering lists. Add a single root word and the filter will automatically conjugate, inflect, and find all possible combinations of the word for you, saving an incredible amount of time and keeping your lists at a small, easily manageable size. CleanSpeak is also fully compatible with any language that uses a Unicode character set including Arabic, Japanese, Chinese and other non-western characters.
Use Whitelist Filtering For Additional Features & Control
Prevent False Positives…
- The pen is mightier than… | As suggested
- Virgin Airlines | Scunthorpe, England
- Assume | Manuscrpit | Glass | molasses
Add Safety For Kids & Teens…
- Enable pre-approved “canned” chat
- Allow only whitelisted words and phrases
- Increase COPPA compliance
CleanSpeak supports even the most complex whitelist configurations and our team can advise you on the most effective way to meet your unique requirements based on our experience working with current clients who have deployed a whitelist-based filtering solution.
Block Personally Identifiable Information & Other Prohibited Content
- Phone Numbers
- Email Addresses
- Social Security Numbers
- Website URLS & Domain Names
- IP Addresses
- Competitors’ Names
CleanSpeak can also find and block any specific words or phrases that may be unique to your application. For instance, in many online games “WTB” stands for “Want To Buy” and can be an indicator of prohibited or unwanted behavior that the game owner may wish to block.
Use Greylisting For Real-Time Behavioral & Contextual Monitoring
Further protect kids & teens and increase COPPA compliance by monitoring your content in real-time to look for inappropriate comments, questions and behavior off all kinds…
Call me. Text me. Email me. Let’s meet… What’s your phone number?
Where are your parents? Are your parents home?
Where do you live? What’s your neighborhood like?
What are you wearing? What do you look like? How old are you?
Used in conjunction with the CleanSpeak Moderation tools, Greylisting can be a simple yet powerful way to immediately identify and take action against users who represent a danger or threat to your community.
Fine Tune The Filter Word-By-Word To Maximize Results
Specify Regional Dialects…
- English: U.S. vs. Great Britain
- Spanish: Spain vs. Mexico
- Any User-Defined Additions
Define Unique Filtering Parameters…
- Set Specific Severity Levels
- Identify Common Variations & Misspellings
- Tag Filter Matches With Violation Type(s)
Customize Response Handling…
- Block, Delete or Replace
- Allow, But Alert
- Take An Automatic Action
Specify Different Filtering Algorithms…
- Exact Match Only
- Distinguishable
- Embeddable or Non-Embeddable
The customization options built directly into CleanSpeak give you the ability to quickly and effectively manage even the most complex filtering requirements. If you are ever stumped by a unique situation, our support team is ready to help you solve the problem with years of hands on, real-world experience.