AGORA BLOG

Ideas for the real-time world

The latest real-time engagement, conversational AI, product, and developer insights from Agora.

WaveSpeedAI founder and CEO Zeyi Cheng on the Convo AI World podcast.

DeveloperOctober 15, 2026

Why the Best AI Model Is No Longer Enough

Learn how WaveSpeedAI helps developers ship generative AI products through unified APIs, flexible model choice, and managed inference infrastructure.

Student speaking with an AI tutor beside a digital whiteboard.

BusinessOctober 12, 2026

How EdTech Platforms Can Scale Student Services with Digital Humans

See how digital humans help educational institutions and EdTech platforms scale advising, tutoring, test prep, and student support around the clock.

Latest articles

DeveloperOctober 15, 2026

Why the Best AI Model Is No Longer Enough

Learn how WaveSpeedAI helps developers ship generative AI products through unified APIs, flexible model choice, and managed inference infrastructure.

BusinessOctober 12, 2026

How EdTech Platforms Can Scale Student Services with Digital Humans

See how digital humans help educational institutions and EdTech platforms scale advising, tutoring, test prep, and student support around the clock.

DeveloperOctober 12, 2026

Keeping AI Agents Accountable in Production

Learn how Trinitite approaches AI agent governance, deterministic inference, observability, and voice AI controls for production systems.

ProductOctober 8, 2026

The Technology Stack Behind AI Avatars in Education

Why Digital Humans in Education? Students do not stop needing support when a class ends. They may want to practice a new language, clarify a course requirement, prepare for an exam

Agora Conversational AI Engine v2.13 connects MLLM custom tool calling to web search, APIs, databases, and other tools.

DeveloperSeptember 28, 2026

Agora’s Conversational AI Engine v2.13: Custom Tool Calling for MLLMs and More

Agora Conversational AI Engine v2.13 adds custom tool calling and MCP support for MLLMs, OpenAI GPT-Live, Smallest AI speech models, and smarter filler words.

DeveloperSeptember 24, 2026

Building AI Avatars Worth Talking To

‍ "Why am I talking to this character? Why do I care?" Jia Shen posed that question while describing AKA Virtual's work on conversational characters. It stayed with us because it u

Woman using a voice AI assistant powered by Agora and Gemini

DeveloperSeptember 22, 2026

A New Generation of Voice: Gemini 3.8 Flash TTS for Voice AI Developers

Explore Gemini 3.8 text-to-speech, expressive voice controls, and what the latest speech generation capabilities could mean for developers building voice AI.

DeveloperSeptember 18, 2026

How Turn-to-Turn Latency Shapes Voice AI

Low latency sounds like the obvious way to improve voice AI. Make transcription faster, and the conversation should feel faster too. AssemblyAI reduced emission latency to roughly

DeveloperSeptember 15, 2026

Gemini 3.8 Live vs. Extended Thinking: Which Is Better for Voice AI

There’s a tendency in voice AI to ask one question: What’s the best model? Google’s new Gemini 3.8 Live models suggest a more useful question: What kind of conversation are you try

DeveloperSeptember 10, 2026

Voice AI That Listens While It Talks: Inside GPT-Live-1

You’re halfway through explaining a problem when the voice assistant starts answering the wrong question. You try to correct it. It keeps talking. By the time you get another chanc

DeveloperSeptember 7, 2026

What Makes an AI Game Host Feel Alive

‍ "They come for the bingo but stay for the connection." Geremie Camara used that line to describe what Live Play Mobile learned from years of featuring live hosts in mobile games.

DeveloperSeptember 2, 2026

Voice AI Gets Real in Production

"The demo worked perfectly." Then a customer interrupted the agent. Another called from a noisy restaurant. A tool call failed, the network slowed down, and an edge case nobody had

BusinessAugust 26, 2026

Gemini Transcribe Is Getting Better at Hearing What Actually Matters

Dive into gemini-3.5-transcribe-live and its new Smart Transcription capabilities. Learn how you can choose between VERBATIM mode to preserve raw conversational speech and SMART mode to automatically clean disfluencies, resolve self-corrections, and format data (like emails, phone numbers, and addresses) into structured inputs for APIs and databases.

DeveloperAugust 26, 2026

Using Gemini 3.5 Transcribe with Agora Conversational AI

Learn how to integrate and use GeminiSTT for converting spoken audio into accurate, real-time text. This guide walks through setting up the API, configuring language parameters, and integrating speech-to-text functionality seamlessly into your software applications.

ProductAugust 24, 2026

Why Enterprise Voice AI Is Harder Than It Looks

Strong voice models still need dependable production systems around them. Learn what enterprise deployments demand beyond realistic speech.

ProductJuly 23, 2026

Introducing the New Agora Console: Build Voice Agents with a Built-in AI Assistant

Build, test, and publish voice agents from one workspace with Concierge, the AI assistant built into the Agora Console.

DeveloperJuly 10, 2026

OpenAI Didn’t Publish GPT-Live’s Latency. So We Measured It

Independent lab benchmarks of GPT-Live reveal end-to-end voice latency, interruption handling, response consistency, and resilience to packet loss on real devices.

DeveloperJuly 2, 2026

Introducing the Agora CLI

A command-line interface for developers and AI coding agents building with Agora. AI coding agents can create apps, but the initial setup still requires human intervention: creatin

DeveloperJuly 1, 2026

Agora Agents SDK: Build Voice Agents in Minutes

Introducing Agora Agents SDK , a typed SDK for building voice agents on Agora's Conversational AI Engine in Python, Node.js, and Go. Voice agents have moved past demos. Teams are p

DeveloperJune 12, 2026

Building Conversational AI Interfaces with Agora Agent UI Kit (Complete Beginner-to-Pro Guide)

The way users interact with applications is changing fast. From static UIs to real-time, conversational, and multimodal experiences , modern apps are expected to feel more human —

DeveloperMay 29, 2026

Voice AI on Android: Beyond Speech-to-Text

Voice AI sounds simple in demos. Tap a mic. Speak. Wait for the AI. Hear a response. But the moment you try to build this inside a real Android app, the problem becomes much deeper

DeveloperMay 21, 2026

Building a Voice AI Agent on Android

A Step-by-Step Deep Dive with Agora and Jetpack Compose Real-time communication (RTC) has become a core requirement for modern apps. From fitness apps with live coaching to gaming

DeveloperMay 15, 2026

Agora Skills: Build Voice AI with Your Coding Agent

Agora Skills gives AI coding agents the current Agora context to build real-time Voice AI apps, tokens, and demos faster.

DeveloperMay 13, 2026

Why Speech Recognition Isn't “Solved”

"Speech recognition is solved." Sure, it is. Until you add background noise. Until your user has an accent. Until two people are speaking at once. I've heard this claim since I sta

Ideas for the real-time world

Featured Blog articles

Why the Best AI Model Is No Longer Enough

How EdTech Platforms Can Scale Student Services with Digital Humans

Why the Best AI Model Is No Longer Enough

How EdTech Platforms Can Scale Student Services with Digital Humans

Keeping AI Agents Accountable in Production

The Technology Stack Behind AI Avatars in Education

Agora’s Conversational AI Engine v2.13: Custom Tool Calling for MLLMs and More

Building AI Avatars Worth Talking To

A New Generation of Voice: Gemini 3.8 Flash TTS for Voice AI Developers

How Turn-to-Turn Latency Shapes Voice AI

Gemini 3.8 Live vs. Extended Thinking: Which Is Better for Voice AI

Voice AI That Listens While It Talks: Inside GPT-Live-1

What Makes an AI Game Host Feel Alive

Voice AI Gets Real in Production

Gemini Transcribe Is Getting Better at Hearing What Actually Matters

Using Gemini 3.5 Transcribe with Agora Conversational AI

Why Enterprise Voice AI Is Harder Than It Looks

Introducing the New Agora Console: Build Voice Agents with a Built-in AI Assistant

OpenAI Didn’t Publish GPT-Live’s Latency. So We Measured It

Introducing the Agora CLI

Agora Agents SDK: Build Voice Agents in Minutes

Building Conversational AI Interfaces with Agora Agent UI Kit (Complete Beginner-to-Pro Guide)

Voice AI on Android: Beyond Speech-to-Text

Building a Voice AI Agent on Android

Agora Skills: Build Voice AI with Your Coding Agent

Why Speech Recognition Isn't “Solved”

Why the Best AI Model Is No Longer Enough