How Kyoso Works

What it is

At the center of Kyoso is an AI agent — your creative partner. You describe what you need, and the agent plans, generates, and edits images and video for you. It picks the right tools and models, remembers your brand guidelines, and can run multi-step workflows on its own.

Everything else in Kyoso exists to support the agent and give it context: boards are where you and the agent work together, brands teach the agent your visual identity, tools give it specialized capabilities, and libraries provide reusable assets and styles.

Key ideas

• Agent — your AI creative partner with two layers: a powerful reasoning engine built by Kyoso, plus a brand-trained layer that adds your organization's visual knowledge on top. It can plan, suggest, run multiple actions, describe images, pull brand guidelines, and more.
• Board — an infinite canvas where you and the agent design together; today the primary surface for interacting with the agent
• Brand — a trained AI identity within the org that holds boards, a trained agent, assets, and image styles; steers the agent's output toward a consistent visual style
• Tools — specialized actions (Edit Image, Upscale, Animate, and more) that the agent uses on your behalf
• Models — a broad selection of AI models for image and video generation; pick one by typing @ in the agent input, or let the agent choose
• Libraries — side-panel collections (styles, assets, mockups, transitions) that you can browse and attach to generations
• Organization — your team's top-level workspace; contains one or more brands and all users

How it fits with the rest of Kyoso

This page is the map. Every section below goes deeper on one of these concepts: see The Agent, The Board, The Dashboard, Brands and the Brand Agent, Libraries, Workflows, and Organizations and Users.

Related

• The Agent
• The Board
• Brands and the Brand Agent
• Libraries