A great idea for audio, documents, images, etc...
I've been using Assistable for my WhatsApp AI agents and ran into an issue with voice notes and media. When a customer sends a voice note, the message arrives with an empty body. The assistant doesn't invoke the LLM, it just repeats the last message (I can see "0 AI responses" in the server logs). This means custom tools can't be triggered either. Attachments also show as "undefined" in the logs. Suggestion: Make the message type and attachment info visible. For example, include the file type (audio, image, document, video), the file URL, and the MIME type. This way we can identify what was sent and users can build custom tools to handle each type through the existing Tool Builder. Would be awesome!! This small change would unlock a lot of use cases (voice transcription, image analysis, document processing) without any major architectural changes. What do you think? @Jorden Williams @Luke Houk @Anthony Castiglia @Jason D @Brandon Duncan @Justin Lee @Assistable Ai @Assistable Team Thanks!