Welcome to the Agora Monthly Digest.
This month’s update highlights our continued advancements in voice-native embodied AI, major RTC and Conversational AI releases, and innovations powering real-time experiences at scale.
Product Updates
Conversational AI Engine v2.5 is now live
Released on March 31, 2026, this update enhances how voice agents are built, managed, and scaled in production.
Key updates:
- Deploy AI agents faster with preset models for ASR, LLM, and TTS
- Track per-turn start, end, and latency events with conversation turn insights
- Dynamically adjust tone, pitch, and speed with real-time TTS updates
- Access new providers including Murf (TTS, Beta) and Anam (Avatar, Beta)
RTC Third-Party Framework SDKs expand on Native 4.6.2
Agora's latest RTC release brings Native SDK 4.6.2 support to Flutter, React Native, Electron, and Unity, expanding real-time performance across more frameworks.
What’s new:
- Cross-platform consistency with a single RTC core
- Faster development across mobile, desktop, and game engines
- Proven ultra-low latency and rich features, now accessible across frameworks
RTC Web SDK v4.24.3 is now available
Agora’s latest Web SDK release improves real-time audio and video performance for browser-based applications.
Key improvements:
- AI Noise Suppression upgrade with low-latency mode
- Improved video experience in H.264 scenarios
- Enhanced Super Clarity for sharper video quality
Global Partnerships
Expanding the real-time AI ecosystem
Agora continues to strengthen its ecosystem to accelerate real-time, conversational, and physical AI capabilities.
- Murf AI integration brings expressive, low-latency TTS to the Conversational AI Engine, enabling more natural, responsive voice agents at scale.
- Seeed Studio collaboration helps power voice-native embodied AI with Reachy Mini and edge AI, enabling real-time, human-like interaction in physical devices.
Success Story
Real-time conversations at global scale
Tandem, a global language exchange platform, uses Agora to power seamless, real-time voice and video experiences across 180+ countries.
The impact:
- 35M+ users supported globally
- 1.7x higher engagement in Language Parties
- 50M+ live visits powered in 2025
- Stable, ultra-low latency communication at scale
Convo AI World Podcast
In this recent episode, Hermes Frangoudis speaks with Ricardo Herreros Symons (Speechmatics) on why 99% accuracy isn’t enough, and what it takes to make speech AI work reliably in real-world conditions.
Happy with Agora? Drop us a quick review on Gartner Peer Insights.
If you have any questions or want help enabling a new feature, we're here to help.
Team Agora
www.agora.io | +1 408 879 5885 | Speak to a team member
LinkedIn | X / Twitter | GitHub
Copyright © 2026 Agora. All rights reserved.
Agora, 2804 Mission College Blvd., Suite 110, Santa Clara, California 95054, USA