Voice chat is the social layer that keeps users engaged and drives deeper connections. Yet toxic behavior in voice communication pushes players away and undermines monetization. Developers need both: voice infrastructure and intelligent moderation. ODIN and Aiba address this challenge by allowing ODIN customers to integrate Amanda, Aiba’s AI-based anti-toxicity solution, into their platforms.
Aiba’s AI-powered safety solution, named Amanda—previously focused on text-based communication—now expands its capabilities to voice data for the first time, protecting communities while unlocking revenue potential across gaming, education and XR. With Amanda, harmful behavior is detected and addressed in near real time, ensuring compliance and privacy by design.
A Win for Communities and Business
Voice builds trust, forges relationships and creates communities. Safer and respectful digital spaces enable even deeper connections. Research shows a fourfold increase in daily sessions and 60% longer average session lengths with proper moderation. Together with Aiba, we’re unlocking this potential to maximize the value of voice chat for everyone.
Phillip Schuster, CEO of 4Players
ODIN gives developers the voice infrastructure to bring their users together. We make sure those communities stay safe and healthy. Together that is a pretty complete picture.
Hege Tokerud, CEO & Co-founder of Aiba
How It Works
ODIN Voice connects players, learners, and communities across platforms. ODIN Cortex, the intelligence layer behind it, captures speech from active voice rooms, transcribes it, and enables Amanda to analyze context, language, and nuances. The results are securely shared with developers via ODIN Cortex, who can then determine appropriate actions such as moderation or automatic mute functions.
The solution ensures:
- Near real-time detection: Toxic behavior, grooming and radicalization in voice chats are flagged within seconds—fast enough to intervene before harmful behavior escalates.
- Context-aware moderation: Amanda reviews conversation context to identify harmful patterns, even if individual lines seem harmless. Difficult cases are routed to human moderators for review.
- Privacy and compliance by design: Voice data is encrypted in transit and at rest, stored only when harmful speech is detected and handled in full compliance with regulatory requirements like GDPR.
- Scalable safety: Applicable to gaming, education, virtual events and any voice-based digital interaction—from small communities to global titles.
- Automated responses: Amanda triggers appropriate actions like warnings, bans or referrals to psychological support. In coordination with ODIN Cortex these actions connect to existing flows, so that the developers always stay in control.
About ODIN
ODIN provides a cross-platform, end-to-end solution from a single source. Powered by a global server network across over 20 locations, businesses worldwide rely on ODIN for seamless digital experiences. The platform empowers developers, enterprises, and public institutions to create secure, scalable and immersive real-time experiences. ODIN is backed by 4Players, a company with more than 25 years of expertise in gaming infrastructure. As part of the Marquardt Group, ODIN delivers technology that bridges the way to success.
About Aiba
Aiba is a Trust and Safety company guided by a single mission: safe digital lives. Its AI moderation platform, Amanda, is built for games and social platforms, detecting harmful behavior in real time across text, image, and voice. Aiba’s work is rooted in over ten years of behavioral science research from the Norwegian University of Science and Technology (NTNU), a foundation that shapes how the company thinks about safety, harm, and the communities it serves.