Images & Vision
sportsclaw can both create images and read them.
Generating graphics
Ask for a graphic and the agent makes one and delivers it right into the chat or saves it locally:
bash
sportsclaw "Make a hype graphic for tonight's Lakers game"In a Discord or Telegram bot, the image is posted straight into the conversation.
Provider requirement
Image generation needs OpenAI, Google, or Azure Foundry as your model provider — Anthropic models can't produce images. Azure Foundry uses the OpenAI-compatible Images API and defaults to gpt-image-2. See Configuration.
Reading images (vision)
Send sportsclaw an image and ask about it — a screenshot of a bracket, a scoreboard, a betting slip — and it reads the picture as part of the question:
- Discord / Telegram — attach an image and ask in the same message.
- CLI — pass an image file path along with your question.
Vision works on any vision-capable model you've connected.