The Ever-Growing Landscape of Content Moderation AI Platforms
If you run a social platform, e-commerce business, or any other online space that includes user-generated content, you understand the importance of content moderation. Luckily, content moderation has come a long way thanks to AI, making it easier and more efficient than ever to implement.
A good question, then, is which content moderation AI platform or company you should choose.
Because there is a broader range of content moderation AI platforms than ever before, you will want to think carefully about your use cases to choose a platform closely aligned with what you need. The most important factors to consider when looking for a platform include:
- What data types does it support? (e.g., video, image, text)
- Does it offer a free option that lets you test out the product before purchasing?
- Does it include any pre-trained models that align with your needs?
- Does it support custom-trained models?
- Does it include human moderation?
In this article, we rounded up 9 high-performing content moderation AI platforms and highlighted their similarities as well as differences so that you can move forward better equipped to choose the right solution for your business.
How do the 9 content moderation AI platforms stack up?
| Data Types | Free option offered? | Pre-trained models available | Custom-trained models supported? | Human moderation included? | |
|---|---|---|---|---|---|
| Nyckel | Text, image, tabular | ✅ Free tier includes 1,000 invokes/month | None, all custom | ✅ Yes | ❌ No |
| Hive Moderation | Text, image, video, audio | ❌ No | Visual moderation, text moderation, intelligent search | ❌ No | ❌ No |
| Imagga | Image | ✅ Free tier includes up to 1,000 API requests | NSFW content moderation | ✅ Yes | ❌ No |
| Sightengine | Text, image, video | ✅ Free tier includes 2,000 invokes/month | Nudity detection, violence detection, face/people detection, image type detection, text moderation, and more | ✅ Yes | ❌ No |
| Membrace | Text, image, video, audio | ✅ Trial available | Image detection, text detection, video detection, audio detection | ❌ No | ✅ Yes |
| Clarifai | Text, image, video, audio | ✅ Free tier includes 1,000 invokes/month | Image recognition, detection, moderation, captioner; multilingual text moderation, text translation, logo detection, OCR | ✅ Yes | ✅ Yes |
| Azure | Text, image, video | 🟨 No, but Microsoft offers $1 per 1,000 invokes | Image moderation, text moderation, video moderation | ❌ No | 🟨 No, but Human Review Tool enables human moderation |
| Spectrum Labs | Text, audio | ❌ No | User retention, text moderation | ✅ Yes | ❌ No |
| WebPurify | Text, video, image | ✅ Free trial available: 5-minutes of video or 100 photos | Photo moderation, profanity and offensive intent detection, video moderation, VR moderation | ✅ Yes | ✅ Yes |
A deeper dive into the platforms and their common use cases
Nyckel
Nyckel, a relative newcomer to the AI content moderation space, is gaining momentum among companies with large, dynamic online communities. Nyckel’s clients include entrepreneurial platform Gust and LGBTQ dating site Taimi.
Nyckel includes classification, search, detection, and OCR for image as well as text content. It also has deep learning options available to users with specific needs or problems that pre-trained models can’t address.
Nyckel offers several tiers of pricing, including a free option at 1,000 invokes per month.
Hive Moderation
Hive Moderation has established a reputation for powering some of the largest social platforms in the world, including Reddit and Vevo.
Hive’s main clients include gaming, streaming, and community platforms, and it has more than 25 pre-trained models across multiple classes, including deep-fake detection. Hive has 15 different APIs across two categories: tagging and search. The company designed its moderation dashboard as an all-in-one content moderation tool to automatically detect and remove problematic content.
Hive does not have pricing readily available on their site, but you can request a demo.
Imagga
Imagga, as its name suggests, offers image-based content solutions. While Imagga offers pre-trained products (notably their Adult Image Content categorizer), it is best known for its customized AI.
Clients use Imagga across a wide range of industries, from real estate to technology to e-commerce. Imagga’s functions allow users to automate the analysis, organization, and searching through unstructured images.
Imagga offers several monthly pricing tiers, including a “test” level that’s free and can process up to 1,000 API requests.
Sightengine
Sightengine specializes in visual content moderation solutions for NSFW and other offensive content.
Sightengine offers off-the-shelf models as well as a higher-tier, customized solution. Its APIs include image and video redaction, moderation, and anonymization as well as text moderation and embedded text detection. Sightengine’s main use cases are dating apps and sites, e-commerce, and content moderation.
Users looking to explore Sightengine can start with a free tier that allows for up to 2,000 invokes per month.
Membrace
Membrace offers content moderation solutions with a focus on e-commerce, dating, and gaming sites.
Membrace’s solutions include 35 classes and 30 models, and the solution combines its machine learning (ML) with human-in-the-loop to moderate over 15 million cases per day. Membrace’s APIs detect problematic content across four types of data (video, image, text, and audio).
Membrace does not offer pricing details on their site, but they allow you to request a trial to test their platform.
Clarifai
Clarifai’s target user base is developers, and the platform caters to a wide range of industries, from the public sector to retail to media and entertainment.
Clarifai has a broad selection of models and workflows that include image recognition, detection, captioning, and moderation as well as text translation, multilingual text moderation, and optical character recognition (OCR). It also offers a custom model-building option for businesses with complex problems.
Clarifai’s “community” tier is free and allows up to 1,000 invokes per month.
Microsoft Azure
Azure’s Content Moderator is one of over 200 products and services that the cloud platform currently offers.
The Content Moderator service offers Moderation APIs that scan and classify video, image, and text data. Users can either use Azure’s built-in list of offensive terms or add their own custom list to block any problematic text. Content Moderator’s concepts also include OCR to extract text from images.
Azure’s Content Moderator technically does not have a free program, but Microsoft offers up to 1,000 transactions for $1.00. It does not include human moderation in its services, but its human review tool allows users to implement their own human-in-the-loop.
Spectrum Labs
Spectrum Labs built its content moderation tool to facilitate better user interaction on gaming, dating, and social platforms. Its solution focuses not only on identifying and eliminating toxic content but on identifying and highlighting positive user behavior.
Spectrum Labs’ text and audio content moderation solutions are multilingual, both with multiple languages built-in as well as the option to add more languages. Its mainstay product, Guardian, moderates users in addition to content by leveraging user reputation scores.
Spectrum Labs integrates into a number of social platforms, including Grindr. Pricing is not readily available on their site, but interested visitors can request a demo.
WebPurify
WebPurify focuses on providing clean content to protect children and ensure that all users have a safe experience online. Their client base ranges from individual bloggers to Fortune 500 companies.
WebPurify offers live moderation services, automated AI, and a hybrid option as well. It allows for video, image, and text data inputs with very simple-to-implement APIs that use HTTP request formats. WebPurify also offers a 15-minute free video trial with no set-up or payment information required.
Choosing a content moderation AI platform to fit your unique needs
While the platforms on our list generally overlap with each other when it comes to the services they offer, most have their own particular strengths and tend towards a specific niche. For example, Web Purify makes content clean and safe for minors, and Spectrum Labs identifies and rewards positive content in addition to moderating problematic content.
While some of these platforms offer customized model learning for complex issues, not all of them do. Your business’s needs will determine whether a pre-trained model or a custom model is better for you.
We hope this article gave you a better understanding of the range of AI content moderation options available – and which is best for your unique situation.
If you’re interested in exploring Nyckel’s solution, check out our content moderation products or sign up for a free account to give it a try for yourself. Reach out to us with questions or for help troubleshooting your specific use cases.