send_conversation_media sends an image, video, or PDF inside an open customer service window. The assistant does not generate the file. You provide a file you already have. Outside the 24-hour window, this send is rejected and the assistant uses an approved template instead.
Ask this
What the assistant calls
1
Read the file rules
get_media_guidelines runs before the upload. HTML, SVG, scripts, and executables are rejected. A partial or cut-off file is rejected and is not a media id.2
Upload
upload_media hosts the file. Pass a public https URL or one complete base64 value up to 5 MB, plus mimeType and filename when you have them.3
Send in the thread
send_conversation_media takes the conversation id, mediaType of image, video, or document, and mediaUrl from the upload or a complete mediaBase64. The caption is optional. The assistant sends it only after you ask.