Getting Started with Tuziyo: AI Image and Video Workflow

Welcome to Tuziyo, a focused workspace for AI image and video generation. You can choose from multiple model families, keep prompts and references in reusable sessions, collect useful outputs in a Library, and arrange visual ideas into ordered shots in Studio.

New accounts receive 10 free generation credits, with no credit card required.

This guide explains the current workflow from the first brief to a reusable set of visual assets.

1. Generate with multiple AI image models

Open the AI Studio, describe the image you want, add references when the selected model supports them, and choose the output controls that fit the brief. Tuziyo keeps the prompt, settings, and results together so you can compare iterations without rebuilding the workflow.

If you are unsure where to start, use the AI image model comparison to compare current resolution, reference, credit, and use-case details.

Different models are useful for different constraints. A quick draft, a product image with strict geometry, a typography-heavy layout, and a detailed campaign scene may each benefit from a different model. Start with the requirement that cannot change, then choose the model around it.

2. Write a brief the model can act on

A useful prompt gives the model a clear job without trying to control every pixel. Describe the request in a stable order:

State the subject and intended outcome.
Define the composition, camera, or layout.
Describe the environment, lighting, materials, and mood.
Identify the details that must remain unchanged.
Add the requested format and meaningful exclusions.

The AI image prompt guide explains this structure in more detail. The prompt examples provide adaptable starting points for products, portraits, characters, interiors, storyboards, and campaigns.

3. Give each reference a role

When the selected model supports reference images, explain what each reference controls. One image might define the exact product, another the lighting, and another the composition or palette. Naming these roles reduces ambiguity and makes comparisons more useful.

Reference capacity varies by model. The model pages document the limits and controls currently available inside Tuziyo rather than repeating a generic provider specification.

4. Preserve the creative trail in sessions

Each generation session keeps the prompt, selected model, reference media, output settings, and generated results together. That makes it possible to return to an earlier direction, compare outputs, or continue from the strongest result without reconstructing the original request.

Use a new session when the creative goal changes materially. Continue an existing session when you are refining the same subject, scene, campaign direction, or shot.

5. Reuse generated assets from the Library

The Library provides a single place to find generated images and videos, mark useful assets, download them, or reuse them in another part of the workflow. Instead of treating every output as a disposable file, keep the candidates that may support future shots, campaigns, or variations.

6. Build a sequence in Studio

Studio turns Library assets into visual projects. Organize images and clips into shots, keep the direction beside each frame, adjust the sequence, and export an ordered MP4 when the project is ready for review.

The goal is not to replace a full professional editor. Studio is a lightweight bridge between isolated generations and a coherent sequence that can continue into a broader production workflow.

Understand the processing boundary

Cloud AI generation sends the prompt, selected references, settings, and required request metadata to the provider for the chosen model. Account, session, task, and output records are stored as described in the Privacy Policy.

Before publishing an output, review identity, product geometry, text, logos, and any other detail that must be accurate. Generative results are candidates, not guarantees.

Ready to test a brief? Open the AI Studio and start with the model that best matches the constraint that matters most.