Saving Images and Documents on Trace for Phoenix Display

Yes, you can save images and documents as part of a trace and have them displayed in Phoenix. For images, include either a base64-encoded image or an image URL in the LLM call; Phoenix will display these in the trace UI. For documents, Phoenix captures and displays retrieved documents in RAG traces, which can be introspected and evaluated within the platform. See the examples and details in the official documentation for multimodal tracing and RAG evaluation support (docs.arize.com multimodal tracing, docs.arize.com RAG evaluation).Would you like a code example or more detail on how to implement this?Sources: - docs.arize.com multimodal tracing

It seems likely that you can include documents (such as retrieved context or input/output text) in a Phoenix trace, and these will be displayed in the UI, especially for RAG use cases. However, uploading arbitrary files (e.g., PDFs, DOCX) as raw attachments to a trace is not explicitly documented—Phoenix expects document content to be included as part of the trace payload, not as separate file uploads. Handling of such content is typically managed by the instrumentation layer, which serializes relevant document data into the trace (docs.arize.com multimodal tracing, RAG evaluation example).Would you like more detail on how to structure document data for traces or on instrumentation best practices?Sources: - docs.arize.com multimodal tracing - RAG evaluation example