03 / MODELS
Images and vision
Attach screenshots and images to Codex on Nebius by selecting a vision-capable model such as Kimi K2.6, using the clipboard shortcut or a file path.
Select a vision-capable model explicitly when launching Codex. The default GLM model is text-only. Image options belong to Codex, not every harness; other agents keep their native attachment controls.
ncodex --model moonshotai/Kimi-K2.6Use Codex's clipboard-image shortcut where your terminal supports it. Normal terminal paste may paste text only; NConnect does not monitor your clipboard. For a reliable file-based alternative, save the image on the machine running Codex:
ncodex --model moonshotai/Kimi-K2.6 --image "/path/to/screenshot.png" "Describe this image"