[frontend] Add video upload support for multimodal models (#826) #830

Closed

sleepy wants to merge 0 commits from fix/video-upload-multimodal-826 into main

sleepy commented

2026-06-03 19:43:10 +02:00

Owner

Adds video upload support to chat for multimodal models with video_input capability.

Video upload button (film icon) visible only when model supports video_input
File picker filtered to mp4/webm, max 50MB, max 60s duration
Thumbnail preview with duration badge, filename, file size, remove button
Sends as video_data + video_format in FormData to backend
Follows exact pattern of direct audio mic button (#825)

Closes #826

Adds video upload support to chat for multimodal models with video_input capability. - Video upload button (film icon) visible only when model supports video_input - File picker filtered to mp4/webm, max 50MB, max 60s duration - Thumbnail preview with duration badge, filename, file size, remove button - Sends as video_data + video_format in FormData to backend - Follows exact pattern of direct audio mic button (#825) Closes #826

sleepy added 1 commit

2026-06-03 19:43:10 +02:00

[frontend] Add video upload support for multimodal models (#826 ) c41e18034f

- Add video upload button (film icon) in chat input, visible only when
  model has video_input capability
- File picker filtered to mp4/webm, with 50MB size limit and 60s
  duration limit using HTML5 video element
- Video preview with thumbnail (generated by seeking to 1s and drawing
  to canvas), duration badge, filename, and file size
- Remove button (X) to cancel before sending
- Base64-encode video and send as video_data + video_format in FormData
- Follows exact pattern established by direct audio (mic button) code
- Listen to odysseus:model-changed event for capability sync

sleepy referenced this pull request from a commit

2026-06-04 00:27:23 +02:00

[llm_core] Eliminate private-import coupling from model_routes and endpoint_resolver (#704, #703)

sleepy closed this pull request

2026-06-04 11:42:14 +02:00