Summarize this article with:
AI-generated media is expanding across every channel, making it harder for teams to verify what they see and hear. Resemble AI is now available on Eden AI, bringing a generative AI security company focused on detecting deepfakes, synthetic media, and manipulation into Eden AI’s unified platform.
Founded in 2019 and trusted by organizations including Netflix, Paramount, and Deutsche Telekom, Resemble AI helps security, fraud, and trust and safety teams assess media authenticity across multiple formats. Developers can now use Resemble AI's deepfake detection - alongside video, image, and audio processing - through Eden AI, with no new integration.
What is Resemble AI?
Resemble AI is a generative AI security company founded in 2019 and based in Mountain View, California. Its mission is to make AI-generated media safer and more trustworthy by helping organizations detect deepfakes, synthetic content, and media manipulation.
At the center of Resemble AI’s detection technology is DETECT-3B Omni. This multimodal model can identify AI-generated and manipulated content across:
- Audio
- Images
- Video
DETECT-3B Omni has been tested against more than 160 generative models and supports over 50 languages and a wide range of media formats. This allows teams to verify different types of content through one detection system instead of relying on separate tools for each format.
Resemble AI’s main differentiator is its understanding of how synthetic media is created. By analyzing the techniques, patterns, and artifacts produced by generative models, its technology can detect manipulated content more reliably.
This approach also enables zero-day coverage, helping identify content created by newly released generative models before large, model-specific detection datasets are available.
Resemble AI complements its detection technology with PerTh watermarking, which helps embed and verify provenance signals in AI-generated media. It also offers identity verification capabilities designed to reduce impersonation and unauthorized use.
Together, these tools help security, fraud, and trust and safety teams detect manipulated content, verify authenticity, and manage the risks associated with generative AI.
Our Interview with Resemble AI’s Head of Developer Relations
To learn more about Resemble AI’s technology, mission, and integration with Eden AI, we spoke with Dev Shah, Head of Developer Relations at Resemble AI. He shared insights into the company’s deepfake detection capabilities, its approach to generative AI security, and what developers can expect next.
Can you introduce your company and its mission?
Resemble AI, founded in 2019 and based in Mountain View, California, is a generative AI security company. Our mission is to make AI-generated media safe and trustworthy by detecting deepfakes and manipulation. Our vision is a world where you can trust what you see and hear.
Can you go into more detail about your offering, your models, and what makes them unique?
We focus on detection and media security. Our DETECT-3B Omni model detects AI-generated and manipulated audio, images, and video. It has been tested against more than 160 generative models, with strong accuracy across over 50 languages and formats.
We complement detection with PerTh watermarking and identity verification. What makes us unique is our deep understanding of how synthetic media is created, which allows us to detect it more reliably, including through zero-day coverage of newly released generative models.
Who are your target users or customers?
Our target users include security, fraud, and trust and safety teams at enterprises and platforms, as well as financial institutions, telecommunications companies, marketplaces, media organizations, and public-sector organizations.
We are trusted by enterprises including Netflix, Paramount, and Deutsche Telekom.
What led you to integrate with Eden AI?
Eden AI gives developers one access point to best-in-class AI providers. Integrating with Eden AI makes our deepfake detection easy to discover, evaluate, and adopt alongside other tools, lowering the barrier for teams that need reliable protection against AI media manipulation.
What’s next for your company?
We are expanding detection coverage across modalities, languages, and emerging generative models. We are also deepening our watermarking and provenance capabilities, building the layer of trust beneath AI-generated media.
Which Resemble AI features are now available on Eden AI?
Resemble AI’s main feature on Eden AI is deepfake and AI-generated content detection, powered by DETECT-3B Omni.
The model is built to detect signals associated with AI-generated or manipulated media, rather than relying only on recognition patterns from a limited set of known generators. Resemble AI’s understanding of how synthetic media is created also supports zero-day coverage, helping teams identify content produced by newly released generative models before large, model-specific detection datasets are available.
This makes the detection API relevant for use cases such as fraud prevention, impersonation detection, content moderation, media verification, and trust and safety workflows. A team can analyze a suspicious voice recording, image, or video without creating a separate integration for each media type.
Deepfake detection is available alongside Resemble AI’s broader audio, image, and video capabilities through the same Eden AI integration.
Resemble AI features available through Eden AI
Audio processing
Developers can access Resemble AI’s audio capabilities for workflows including speech-to-text and text-to-speech. These tools can support transcription, voice interfaces, media applications, automated content production, and accessibility features.
Audio processing can also be combined with deepfake detection. For example, a platform can transcribe an uploaded recording while also checking whether the voice content may have been generated or manipulated.
Image processing
Resemble AI’s image-processing capabilities allow applications to analyze visual content through the Eden AI API. When paired with DETECT-3B Omni, teams can add authenticity checks to image-upload workflows, moderation systems, marketplaces, or verification tools.
This helps developers process an image and evaluate potential synthetic manipulation without maintaining a separate provider integration.
Video processing
Video processing extends the same approach to moving media. Applications can analyze video files and use Resemble AI’s detection technology to identify indicators of AI generation or manipulation across the visual and audio components of the content.
By making detection, audio, image, and video processing available through Eden AI, developers can add Resemble AI to an existing AI workflow without implementing and maintaining another standalone API connection.
Which Resemble AI Feature Should You Use?
The right Resemble AI feature depends on whether your workflow needs to detect manipulated media, verify content provenance, or process audio.
For fraud prevention, KYC, and identity-verification workflows
Start with DETECT-3B Omni. The model analyzes audio, images, and video for signs of AI generation or manipulation. It is the most relevant option when users submit identity documents, profile images, voice recordings, or verification videos that could have been altered or synthetically generated.
Executive-impersonation and trust and safety workflows often require a combination of audio and video detection. For example, a security team may need to review a suspicious voice message, video call recording, or social media clip that appears to feature an executive, employee, or public figure. Using detection across both modalities helps teams evaluate the full media asset instead of checking only the audio or visual layer.
For content provenance and compliance
Use PerTh watermarking. Watermarking helps organizations embed and verify provenance signals in AI-generated media. It is useful when teams need to document where content came from, distinguish authorized synthetic media from unverified content, or support internal governance requirements.
For applications focused on transcription or voice output rather than media verification, use Resemble AI’s audio capabilities, including speech-to-text and text-to-speech. These features support workflows such as meeting transcription, accessibility, voice interfaces, automated narration, and customer-service applications.
Some products may need more than one feature. A marketplace, for example, could use speech-to-text to transcribe a submitted recording and DETECT-3B Omni to check whether the voice was artificially generated. A media platform could apply watermarking to approved synthetic content while using deepfake detection to review external uploads.
How to access Resemble AI on Eden AI
You can access Resemble AI through Eden AI in a few steps:
- Create a free Eden AI account.
- Generate your API key from the Eden AI dashboard.
- Select Resemble AI as the provider for deepfake detection or another supported feature.
- Send your request through the Eden AI API.
Because Resemble AI is available through Eden AI’s unified integration, you do not need to create or maintain a separate provider connection. You can also combine its deepfake detection capabilities with audio, image, and video processing in the same application.
curl -X POST "https://api.edenai.run/v3/universal-ai/async" \
-H "Authorization: Bearer YOUR_EDEN_AI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "video/deepfake_detection_async/resemble",
"input": {
"file": "https://example.com/video.mp4"
}
}'
The response will contain Resemble AI’s analysis, including the available indicators or confidence scores for AI-generated or manipulated content.
Conclusion
Resemble AI is now live on Eden AI, giving developers one API for deepfake and AI-generated content detection alongside audio, image, and video processing. With DETECT-3B Omni, teams can add a trust layer to fraud prevention, identity verification, content moderation, and media-authenticity workflows. Resemble AI’s broader audio and provenance capabilities can be added through the same Eden AI integration, without building and maintaining a separate provider connection.
.jpg)
.png)
.png)
