OJO PlayBook

How to Use OJO: A Step-by-Step Guide

A practical walkthrough for building an Agent Team, choosing the right workflow, refining a prototype, and exporting the result with OJO.

  • Getting started
  • Agent Teams
  • Prototyping
OJO product design workflow connecting an idea, specialist agents, an editable canvas, and an interactive prototype

Quick answer: To use OJO, define a real product or design goal, assemble an Agent Team, add the Skills needed for the current stage, generate a first result on the Canvas, and improve it through specific feedback. Validate the final UI, website, or interactive prototype before sharing or implementation.

The most reliable way to use OJO is to bring a real requirement or existing asset, choose an Agent Team for the scenario, add only the Skills needed, create a first prototype, and improve it through specific feedback on the Canvas.

OJO is designed as an AI Design Agent Team Workspace, not a one-click page generator. The quality of the result depends on how clearly the goal, audience, constraints, and desired deliverable are defined.

1. Start with context, not a one-line request

Explain who you are, what you need to deliver, who the audience is, where the result will be used, and what success looks like. Include existing copy, screenshots, URLs, research, or brand material when they matter.

A useful starting brief includes:

  • project and product type;
  • target user and core problem;
  • desired outcome or conversion action;
  • required pages or product states;
  • references and what to learn from them;
  • visual and technical constraints;
  • delivery format and deadline.

If the product idea is still vague, ask OJO to clarify assumptions and propose a focused brief before generating the interface.

2. Assemble the right Agent Team

Choose the team that fits the scenario rather than adding every available role. A website redesign, product prototype, campaign asset, and design-to-code task require different attention.

An Agent Team can keep product thinking, visual exploration, UI creation, and review connected. You remain responsible for the goal and final judgment; the agents help explore and execute the work.

3. Add Skills for the current stage

Skills provide focused methods and constraints. Add them when the project needs stronger product thinking, page hierarchy, typography, layout, motion, performance, or design review.

More Skills do not automatically produce a better result. Use the smallest set that matches the current stage, then change the set as the work evolves.

4. Choose a model and working mode

Use the model and mode that fit the task and review cycle. A visually demanding direction may need a stronger visual model; repeated revisions may favor a faster option. Exploratory work benefits from a mode that reasons through choices, while a tightly specified task can use a more direct execution mode.

The important point is to make the trade-off intentionally rather than using the most expensive option for every step.

5. Generate the first prototype on the Canvas

Ask for a complete, reviewable direction rather than an isolated hero image. For product UI, include the important states and transitions. For a landing page, include the information hierarchy, proof, and conversion path. For a redesign, identify what must be preserved.

After generation, inspect the relevant output format and any required device states. Check whether the result reflects the brief before asking for visual polish; do not assume that a preview alone proves responsive implementation.

6. Give targeted feedback

Click or reference the specific element that needs work. State the problem, desired change, and reason.

Weak feedback: “Make this look more premium.”

Useful feedback: “Reduce the visual weight of the feature cards, give the customer proof more space, and keep the demo CTA visible after the pricing comparison.”

Specific feedback lets the Agent Team revise the existing direction without discarding useful decisions.

7. Validate, share, or continue to implementation

Before delivery, review content accuracy, responsive behavior, important states, accessibility, and the core user path. Then deploy the prototype, export a supported format such as HTML, PPT, PDF, PNG, or JPG, or continue into a design-to-code workflow. A coding workflow can use OJO’s CLI to connect project assets with tools such as Codex, Claude Code, or Cursor; generated code still requires engineering review.

For an end-to-end example, follow the fictional idea-to-interactive-prototype walkthrough.

Common mistakes when using OJO

OJO works best as an iterative AI design workflow: frame the goal, direct the team, inspect the result, and refine the decisions that matter.

Frequently asked questions about using OJO

What can you create with OJO?

OJO can support product thinking, structured requirements, UI directions, landing pages, website prototypes, interactive product prototypes, and campaign assets. It can deploy a result, export supported formats, or pass assets into a coding workflow; exact release capabilities should still be checked before publication.

Is OJO an AI UI generator?

OJO can generate UI, but it is positioned as an AI Design Agent Team Workspace. Agent Teams, Skills, shared context, and the editable Canvas support the product decisions and revisions that happen before and after the first screen.

What information should a first OJO prompt include?

Include the audience, problem, desired outcome, required deliverable, existing assets, references, constraints, and what success looks like. Ask OJO to flag assumptions when important information is missing.

What current product limitations should users know?

Current product limits include the following: chat does not accept video files directly, so convert a video to a URL before inserting it into a prototype. Direct font-file upload is not supported. OJO CLI can pass OJO assets to coding tools, but it cannot currently upload or update a local project back into OJO. These limits may change as the product evolves.

Start with a real task in OJO

The fastest way to judge whether OJO fits your workflow is to use a real brief, reference, or existing page—not a generic test prompt. You can review the resulting structure and prototype before deciding whether to deploy, export, or continue into code. Open OJO and start your first Agent Team project.

OJO

Turn your next idea into a working product

Build research-backed UI, landing pages, and interactive prototypes with an OJO Agent Team.

Start with OJO