Images

OpenDunes supports image inputs on vision-capable models (pass them inline in messages[].content), and image generation from text on dedicated image models.

Image inputs follow the OpenAI chat-completions content-part schema, so an OpenAI-compatible Go client pointed at https://opendunes.com/v1 sends them through its chat service — check your installed SDK version's own documentation for the exact content-part helpers, since they change between releases. The wire surface below is the contract any Go HTTP client can target directly.

Image inputs (vision)#

Send a user message whose content is an array mixing an image_url part (an HTTPS URL or a data: base64 URL) and a text part:

curl https://opendunes.com/v1/chat/completions \
  -H "Authorization: Bearer $OPENDUNES_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/gemini-2.5-pro",
    "messages": [
      { "role": "user", "content": [
        { "type": "image_url", "image_url": { "url": "https://upload.wikimedia.org/wikipedia/commons/thumb/7/7f/Sahara_desert.jpg/800px-Sahara_desert.jpg" } },
        { "type": "text", "text": "Describe this landscape." }
      ] }
    ]
  }'

The reply comes back in choices[0].message.content, exactly like a text completion.

Image generation#

Generate images from text prompts via POST /v1/images/generations — see Generations for the Go example.

Supported vision models#

Filter the catalog for vision-capable models: GET /api/models?capability=vision. Any model with the vision capability accepts image content parts in the messages array.