Image input

Attach screenshots, photos, designs, charts, or scans so EvoX can use them as visual context for the current task.

Add an image

Choose an image from the add menu or drag it into the composer. An image belongs to the current conversation or project context; it does not automatically appear in another conversation with the same project name.

Prompt around the image

State what the image shows, the region to inspect, the required output, and what must not change. Label multiple images and define the comparison order.

Image 1 is the current screen and image 2 is the target. Fix spacing, typography, and alignment only. Do not change behavior. Verify desktop and mobile with new screenshots.

Treat unreadable text, occlusion, and low-resolution detail as uncertain instead of guessing.

Choose the right capability

• Use image input to inspect, compare, or extract visual information.
• Use image generation to create or edit an image.
• Use the built-in browser to open and inspect web pages.
• Use Attach an app to select one open window and capture its latest state at send time.

Privacy

Crop accounts, notifications, local paths, customer information, and verification codes. Parsing scope depends on the active model, plugin, and attachment pipeline; a successful upload does not prove every detail was read.

EvoX Docs · Features · Capabilities