Cookie Preferences
close

By clicking “Accept”, you agree to the storing of cookies on your device to enhance site navigation, analyze site usage, and assist in our marketing efforts. View our Privacy Policy for more information.

Close icon
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Our product

Our context aware AI understands the nuances of each piece of content.

Unitary's technology combines the latest advances in computer vision with natural language processing techniques. We weave data points into a knowledge graph to create context aware AI that automatically extracts meaning from video, text and audio.
Happy emoji icon
Scalability

Our platform classifies over 25,000 frames of video per second. That's three billion images per day.

Flag icon
Accuracy

Our data is regularly tested on various labeled datasets. Plus, we always interrogate bias in our dataset and resulting models.

Strong arm icon
Flexibility

Our custom AI is delivered at speeds that work for you and offers capabilities in 6+ languages.

Warning icon

How it works

1
Tell us about your current policies and safety challenges.
2
We create a custom machine learning model.
3
We'll rigorously test your custom model against real world scenarios.
4
We'll seamlessly integrate our AI into your workflow, using scalable APIs.
1
Tell us about your current policies and safety challenges.
2
We create a custom machine learning model.
3
We'll rigorously test your custom model against real world scenarios.
4
We'll seamlessly integrate our AI into your workflow, using scalable APIs.

Unitary's technology

Our platform simultaneously analyses multiple signals via multimodal models. It allows us to take into account the context behind each piece of content, and build up a holistic view of its meaning.
1
Upload content
Upload all your content via our API.
Robot API illustration
2
API does its magic
We simultaneously extract visual content, audio, captions and OCR.
Video files illustration
visual content
Text, title, descriptions, captions
Audio, voice, soundtrack
OCR, on-screen text, closed captions
3
Content in context
These signals are inputted into Unitary's multimodal models to interpret the meaning of the content in context.
Robot scanning illustration
Comments icon
comments
Play video icon
captions
Audio icon
audio
Image icon
video stills
4
We deliver the output
We deliver the output, which is customised content classification.
Illegal drugs, alcohol & Tobacco
NSFW
Adult & Explicit sexual content
Arms & Ammunition
Crime & Harmful Acts
Terrorism
For Advertisers: Brand Safety

Brand Safety
in line with GARM

Adult & Explicit Sexual Content
Debated Sensitive Social Issue
Arms & Ammunition
Crime & Harmful acts
Online piracy
Death, Injury or Military Conflict
Spam & Malware
Misinformation
Hate speech & acts of aggression
Obscenity and Profanity
Debated Sensitive
Illegal drugs & substances
Terrorism

Our Brand Safety product ensures that your brand - and users - are kept safe. A single piece of content can be disastrous for a brand, and traumatic for a user.

We use the GARM standard for advertising and media, to create a safer digital media environment for everyone.

  • Use a tool that understands the nuances in your safety policies.
  • Streamline and automate your moderation.
  • Protect your brand’s reputation.
For Platforms: Trust & Safety

Content analysis tailored to your platforms

Our Trust & Safety tool is designed to accommodate any platform.

Every community is unique, and has its own definition of safety - so we’ve made sure that you can incorporate nuances into your content moderation. We’ll work with you to create custom categories depending on the specific needs of your platform.

  • Use a tool that understands the nuances in your safety policies.
  • Remove harmful content at the point of upload.
  • Use AI to reduce the risk to your team’s mental health.
Detoxify: Text-Based Trust & Safety

Automatically flag
and remove harmful comments

Our mission at Unitary is to create a safer internet for everyone. We created Detoxify with this in mind.

Detoxify is a simple open source Python library, designed to identify comments containing toxic language. It functions in six languages, and is able to incorporate bias into its analysis.

I hate you
Compute
Computation time on cpu: cached
0.951
toxic
0.138
insult
0.039
threat
0.024
obscene
0.021
identity_hate
0.008
severe_toxic

Drop us a message to learn more - we’d love to hear from you

We partner with game-changing companies and individuals looking for ways to protect their communities from online harm and safety-proof their brands. Get in touch.