AI Look Training in Evoto helps you train your own personal color style. By learning from your presets, XMP data, or edited image sets, Evoto can generate a reusable AI Look that reflects your editing preferences and helps keep color results more consistent across future images and projects.

Personalized AI Look now starts from the current Create AI Look workflow. In AI Look > Train Your AI Look, Color Train opens with two creation options: Lite Personal AI Look and Personal AI Look.

Create AI Look

When you enter the Color Train flow, the first screen shows only these two options:

• Lite Personal AI Look: Create a Look from a single .xmp file or an Evoto color preset.
• Personal AI Look: Create a Look from a larger edited image set and continue with image-type settings and source selection.

Lite Personal AI Look

Choose Lite Personal AI Look when you want to build a Look from an existing color style instead of preparing a large training set.

Basic Setup

• Look Name: Enter a name for the new Look.
• Training source: The Lite workflow supports two source options in the current UI.

Source Options

• Upload .XMP: Train the Look using color data from an external .xmp file.
• Select Evoto Color Preset: Train the Look using an Evoto preset.

Notes

• Only .xmp files and Evoto personal color presets are supported.
• AI Color Preset is not supported.
• The preset must include essential color data such as white balance, exposure, and tone.
• Presets with local masks may produce less accurate results.
• If the XMP references a profile, enable Auto-Import Profiles from XMP.

After the source is selected, Evoto generates the Look and sends you to the preview step so you can confirm and fine-tune the result before use.

Personal AI Look

Choose Personal AI Look when you want to train from a larger edited image set. This flow is organized into Basic, Settings, and Select Image Source.

Step 1: Basic

• Look Name: Enter a name for the Look.
• Main image type: Choose the type of images you mainly upload, such as Wedding, Portrait, or Family.

Step 2: Settings

• Look is for: Choose Color or Black & White.
• Training image type: Choose RAW, JPEG/TIFF, or Both.
• Image requirements: RAW requires 200+ images, JPEG/TIFF requires 200+ images, and Both requires 200+ images for each format.
• Availability note: Your current settings affect which image sources are available in the next step.

Step 3: Select Image Source

The source options in this step depend on the image type selected in Settings.

If RAW Is Selected

• Lightroom Classic: Import edited images from Lightroom Classic.
• Evoto: Use edited images from Evoto projects or collections.
• Local Folder (Raw + XMP): This is the additional RAW + XMP workflow. It imports original RAW files together with XMP data that contains required color parameters such as white balance, exposure, and tone.

Note

• The Local Folder (Raw + XMP) option is available when creating a Color Look for RAW images.
• You can select a folder or drag and drop files into the training flow.
• Evoto detects usable images automatically before upload.
• You can optionally enable AI Smart Selection before upload.
• Before training starts, Evoto reports the reason and count for unusable files.
• If the XMP references a profile, enable Auto-Import Profiles from XMP.

If JPEG/TIFF Is Selected

• Lightroom Classic: Import edited images from Lightroom Classic.
• Evoto: Use edited images from Evoto projects or collections.

If Both Is Selected

• Lightroom Classic: Import the training set from Lightroom Classic.
• Evoto: Use images from Evoto projects or collections.

Image Import Details

• Lightroom Classic: Supports .lrcat file selection, drag and drop, and browser import. You can filter by folder, collection, star rating, flag, or color label.
• Evoto Project: Select images from projects or collections. Cloud projects must be downloaded before they can be used for training.
• AI Smart Selection: Automatically filters images with better diversity and color consistency to improve training quality. This is available when at least 50 images are available.

Preview & Adjustment

Import up to 5 images for Look preview, then adjust the result before training.

• Basic: Includes AI Correction and Fixed Value modes.
• Curve, HSL, Color Grading, Color Calibration: Available for fine adjustment. Color Calibration works in AI Correction mode only.
• AI Correction: Applies offset correction based on AI-generated values.
• Fixed Value: Directly replaces the AI-generated value.

After Training

After a Look is trained, you can manage it from the Look panel, duplicate it, or track related tasks in the global progress manager.

Look Panel

• Status: Looks can appear as Unready, Ready to Use, Ready to Use (Training), or Locked after a free quota is exceeded following a downgrade.
• Format matching: RAW and Non-RAW Looks can only be applied to matching image formats.
• Preset saving: Looks can be saved as presets. When presets are shared, the Look effect is masked because Look sharing is not supported directly.

Look Duplicate

• Create a copy of a trained Ready Look quickly.
• Use duplication to preserve the original result while refining a new style variant.

Global Progress Manager

• After Look training starts, the task appears in the global progress manager.
• You can review task status there and use available quick management actions.
• Only Waiting and Uploading states support cancellation. Tasks already in Training cannot be canceled.