[3.13] Update AI Proxy documentation for audio, video, and image editing/generation and rerank across providers

[tomek-labuk]

Jobs to be done

In 3.13 we are expanding coverage beyond text to multimodal AI capabilities. Update ai-proxy and ai-proxy-advanced provider feature matrix and associated examples to cover:

• Audio input and audio output support across providers (AWS Bedrock, Gemini, Azure)
• Video support notes and examples where applicable (upload and processing flows)
• Image generation and image understanding support across providers
• Gemini rerank endpoint support and example

Implementation tickets:

• https://konghq.atlassian.net/browse/AG-459
• https://konghq.atlassian.net/browse/AG-362

Definition of Done

• Provider capability matrix updated in

  • AI Proxy Advanced documentation
  • AI Proxy documentation

• New plugin examples include:

  • Audio-to-audio example (Gemini)
  • Audio input and transcription example (Bedrock, Azure)
  • Image input example (Gemini, Azure)
  • Image generation example (Bedrock, Azure)
  • Video usage notes and supported paths
  • Gemini rerank request and response example

• Provider differences and API capabilities clearly documented

  • Format requirements (audio types, image formats, etc.)
  • Streaming vs non-streaming availability
  • Model naming examples

Information

Person of contact: Wangchong

Size

L