Background
Get Started

Insights

Field notes on real-time AI, agents, and building with Ojin.

What a Human AI Agent Is, and When It Beats Recorded Video

Featured

A human AI agent is a face and voice you can talk to live, not a recorded clip. How the real-time kind works, where it wins, and how to build one.

Read article

Why We Bet on Real-Time, Not Recorded

Up next

The thesis behind Ojin, and what changes when latency disappears. Coming soon.

When latency disappears, video stops being a recording and starts being a conversation.

Ojin Insights

Coming soon

Designing a Conversational AI Agent With Real Presence

Modular agents with lifelike voice and face — the architecture behind a natural back-and-forth.

01

The AI Sales Agent That Actually Closes

Where real-time agents outperform forms and chatbots across the funnel.

02

Building on the Real-Time Agent API

Speech-to-video in under a second: the API surface and the latency budget.

03

Guides

Section

How to build and ship human AI agents, step by step.

Engineering

Section

Latency budgets, APIs, and the infrastructure behind real time.

Playbooks

Section

What works in production, from sales floors to support desks.