Attachments

Add images, screenshots, videos, or dictated text when showing Ziva is easier than describing the problem.

Images and screenshots

1. Paste an image or screenshot into the composer, or open Attach file and choose Add photos or files.
2. Add a question that tells Ziva what to inspect.
3. Send the message normally.

You can also ask Ziva to take a screenshot of the whole Godot editor or the 2D, 3D, or Ziva view. Images are resized to fit within 2000 x 2000 pixels and the image size limit when possible; very large images and GIFs that would need resizing are rejected.

Videos

1. Open Attach file and choose Add photos or files.
2. Select an MP4, MOV, or WebM video up to 100 MB.
3. Add your question and send. If the upload is still running, the message waits for it to finish.

Use a video-capable model. If the selected model cannot watch videos, Ziva warns that the attachment will be skipped. Image understanding also depends on the selected model.

Voice dictation

1. Click the microphone button labeled Start recording.
2. Speak, then click Stop recording.
3. Review the transcript added to the composer, edit it if needed, and send it.

A recording stops automatically after one minute.

Privacy

• Images and videos upload to Ziva’s private storage and are sent to an AI model for analysis when needed.
• Voice recordings are sent to Ziva’s transcription service to turn them into text.

Troubleshooting

The microphone does not start. Allow microphone access for the editor, then try Start recording again.

An attachment is rejected. Use a smaller image or video. For a large animated GIF, use a smaller GIF or a static image.