VideoSDK's Blogs - VideoSDK

Introducing "NAMO" Real-Time Speech AI Model: On-Device & Hybrid Cloud

NAMO real-time speech AI model is VideoSDK's open-source SSLM with hybrid on-device and cloud inference for low-latency, privacy-first voice agents. Compare architectures, integrate via Agent SDK, and avoid production pitfalls.

Build an Online Classroom App with React

A step-by-step guide to building a real online classroom in React, with a shared whiteboard, live video, and role-based permissions, powered by VideoSDK.

Build ESP32 Video & Audio Calling With VideoSDK

Build real-time ESP32 video calling with VideoSDK's IoT SDK. Join a browser and an ESP32-S3 into the same WebRTC room

How to Build an Open-Source Zoom Clone with Claude Code + VideoSDK

Build an open-source Zoom clone in about 10 minutes. Clone an MIT repo, add free VideoSDK keys, customize it with Claude Code, and deploy.

How to Build an Open-Source Google Meet Clone with Claude Code + VideoSDK

Build an open-source Google Meet clone in about 10 minutes. Clone an MIT repo, add free VideoSDK keys, customize it with Claude Code, and deploy. Beginner friendly.

Echo Models: Linguistic Turn Detection for Voice Agents

Echo Turn Detection is a real-time timing layer that transitions voice agents from rigid, walkie-talkie-like interfaces to natural, human-like dialogue. By analysing linguistic patterns rather than simple silence timeouts, Echo delivers up to 96.2% turn-taking accuracy and 25x fewer missed turn endings, allowing users to pause, hesitate, interrupt, and backchannel naturally. The Turn-Taking Challenge In a voice agent pipeline, speech recognition gets most of the attention. But knowing what th

Introducing VideoSDK Warm Transfer: Escalate to a Human - Briefed, Before They Take the Call

Escalate from your AI voice agent to a human the right way with VideoSDK Warm Transfer—brief the supervisor with an AI summary before the caller connects.

Introducing VideoSDK Call Transfer: Let Your AI Agent Hand Off Live Calls - No Hang-Up, No Redial

Add human escalation to your AI voice agent with VideoSDK Call Transfer—forward a live SIP call to any number with no disconnect and no redial.

Introducing the VideoSDK Fallback Adapter: Keep Voice Agents Talking When Providers Fail

Keep AI voice agents talking when a provider fails. The Fallback Adapter adds automatic STT, LLM, and TTS failover—on error or latency—with mid-session recovery.