How Glossi works
The pieces of a Glossi workflow, from a 3D model to a reusable template, a finished render and a delivered output.
Glossi keeps the product and the scene behind every image, so the next render starts from approved work instead of from scratch.
The pieces
| In Glossi | What it holds |
|---|---|
| Workspace | Members, assets, projects, usage and settings |
| Model | The 3D product: its parts, materials, variants and versions |
| Project | The editable scene you open in Studio: products, lights, background, props, cameras and motion |
| Shot or clip | One saved camera view for an image, or one saved camera move for a video |
| Template | A versioned copy of a project's setup, applied to other products |
| Render | A finished image or video, with any AI edits as new versions of it |
The workflow
Bring in the product
Upload a 3D file, generate a starting model from photos or technical drawings, or start from a sample. Glossi converts the file and marks the model ready when Studio can use it. Check its scale, parts and materials before you build several projects on it.
See Upload a model and Prepare a model.
Set it up
Create variants for the product's finishes, apply materials across models, and add logos to decal slots. See Variants.
Direct the scene in Studio
Opening a project starts an Unreal Engine session and streams it to your browser. Choose materials, light the scene, pick a background, add props, and save shots and clips. Changes are saved to the project.
See Studio tour.
Save the setup as a template
Once a project produces approved work, save it as a template. Each save is a new version, and a project can stay on the version it was built from. See Templates.
Produce
| Use | When |
|---|---|
| Studio | You are still directing the scene. See Render. |
| Render products | A template is ready and you want it on more products. See Render products. |
| Automations (Enterprise) | A folder or event should start a known workflow. See Automations. |
| API or agents (Enterprise) | Your own system or an AI agent should drive the work. See API and Agents. |
Review and deliver
Finished renders land in Renders. Comment on them, edit them with AI, and download them. Deliver through connected folders, webhooks or a Showroom. See Render library.
Where AI fits
AI helps at a few steps. The 3D model stays the source of truth.
- Generate a starting model from photos or technical drawings.
- Prepare a model: split it into parts, group parts by finish, unwrap UVs, assign materials.
- Generate materials.
- Change the scene in plain language with the Assistant.
- Edit a render with a prompt, saved AI Style, reference image, annotation or layout guide. Each edit is a new version and the original stays.
Generated output can change product details. Compare each edit with the original before you deliver it. See Edit with AI.
What you need
- A current desktop browser and a stable internet connection.
- A 3D model, or clear product photos or technical drawings.
- A workspace with usage left for the work you plan.
Geometry editing, simulation and custom shaders stay in your 3D tool. See From your 3D tool.