feat(sdk/go): add support for image inputs in ai calls

[BenjaminGoehry]

Closes #99

Changes

Modified Files

• agent_test.go – Updated tests to reflect the new struct.
• client.go – Adapted request handling for OpenRouter compatibility.
• client_test.go – Updated tests to use the new struct.
• config.go – Added transformForOpenRouter to map OpenAI-style JSON to OpenRouter format.
• request.go –
  • Updated structs for OpenAI-style image inputs.
  • Added WithImageFile, WithImageURL, and WithImageBytes for image support.
• request_test.go – Added tests for the new image-related functions.
• README.md – Updated documentation for the new options.

New File

• multimodal.go: Added helper functions such as detectMIMEType.

Features

✅ Image Support:

• Send images via file path, URL, or raw bytes
• Automatic MIME type detection for common formats (JPEG, PNG, etc.)

✅ Error Handling:

• Skip empty inputs for WithImageBytes and WithAudioBytes
• Return errors for non-existent files

[github-actions]

Performance

SDK	Memory	Δ	Latency	Δ	Tests	Status
Go	225 B	-20%	0.78 µs	-22%	✓	✓

✓ No regressions detected

[AbirAbbas]

@BenjaminGoehry just checking to see if I should review this or not (seems to be in draft)

[BenjaminGoehry]

Hi @AbirAbbas ,
I’ve updated all the files. The image input now works fully with the OpenRouter API. I wasn’t able to get audio files working yet—I don’t have an OpenAI key to test, but I implemented the mapping based on OpenAI docs and OpenRouter docs.

The changes for images were more significant than expected, so I think the PR is ready to review as-is. I can follow up with a separate PR for audio input later if that’s ok.

Thanks!