Adding Rendered Images to Phoenix for Mermaid Diagrams

Is it possible to add rendered images to phoenix? For example if I'm generating a mermaid diagram I'd love to see it rendered in-trace.

· Sorted by Oldest

RunLLM
APP
·
Yes, Phoenix supports displaying images in traces, including those encoded in base64 or provided as image URLs. According to Capture Multimodal Traces, you can include a base64 UTF-8 encoded image or an image URL in the call made to your LLM, and Phoenix will render these images in the trace UI. Would you like more details on how to implement this? Sources: - Capture Multimodal Traces
💡 Hint: Mention RunLLM in the thread for followups.
Kashyap C.
·
So this is in the call made to the llm. what I'm suggesting is rendering the generated mermaid diagram in the output RunLLM
Mikyo
·
Looks like there's a mermaid plugin for remark - if you file us an issue we can take a look at what it would take to add! https://www.npmjs.com/package/remark-mermaid-plugin

RunLLM
APP
·
Yes, Phoenix supports displaying images in traces, including those encoded in base64 or provided as image URLs. According to Capture Multimodal Traces, you can include a base64 UTF-8 encoded image or an image URL in the call made to your LLM, and Phoenix will render these images in the trace UI. Would you like more details on how to implement this? Sources: - Capture Multimodal Traces
💡 Hint: Mention RunLLM in the thread for followups.
Kashyap C.
·
So this is in the call made to the llm. what I'm suggesting is rendering the generated mermaid diagram in the output RunLLM
Mikyo
·
Looks like there's a mermaid plugin for remark - if you file us an issue we can take a look at what it would take to add! https://www.npmjs.com/package/remark-mermaid-plugin