Images
Image Input
Send images to vision-capable models through supported protocol formats.
Image input is different from image generation. It sends an image to a model for understanding, extraction, or reasoning.
| Protocol | Typical image format | Endpoint |
|---|---|---|
| Anthropic-compatible | content block with type: "image" and base64 source | /v1/messages |
| OpenAI-compatible | image_url or input_image, depending on API style | /v1/chat/completions or /v1/responses |
| Gemini native | inlineData with mimeType and base64 data | /v1beta/models/{model}:generateContent |
Image input support depends on the selected model. If a model rejects image input, choose a vision-capable model in the same provider group.
AnyRoute Docs