You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 2eb80dd
Browse filesBrowse the repository at this point in the historyBrowse files
authored
feat(media): add Higgsfield to video and image generators, allow typed models (#9756)
* feat(media): add Higgsfield to video and image generators, allow typed models
- Add Higgsfield as a provider in the Video Generator (Kling 3.0, Kling 2.5 Turbo, Seedance 2.0, MiniMax H3, Hailuo 2.3) and Image Generator (SOUL, SOUL V2, SOUL Cinema, Z-Image Turbo)
- Switch the video and image model fields to comboboxes so users can type a model the catalog does not list
- Send typed model IDs with the prompt only; validate them as safe model names or endpoint paths
- Keep retired model errors, reject stale cross-provider values, and require a user Fal.ai key for custom Fal endpoints
* fix(media): trim typed model IDs, guard reference images, restore default after provider switch
- trim `model` at the image and video input schemas so a padded catalog ID still matches its catalog entry
- reject `visualReference` for typed direct-video models, which only send the prompt
- refill the combobox default after a provider switch even if the user had cleared the field
- update the image operations test to the typed-model behavior
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(media): test typed image models at the request builder, document Higgsfield helpers
- move the typed-model checks from mocked fetch calls to buildImageGenerationRequest, covering padded catalog and retired IDs
- add TSDoc to higgsfieldStatusUrl and higgsfieldFailureMessage
- say in the combobox comments that a cleared required field gets its default back when reopened
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* refactor(higgsfield): share one API client between the block tools and media generators
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Waleed Latif <walif6@gmail.com>
Copy file name to clipboardExpand all lines: apps/sim/app/workspace/[workspaceId]/w/[workflowId]/components/panel/components/editor/components/sub-block/components/combobox/combobox.tsx
description: 'Generate videos from text using AI',
1917
1933
longDescription:
1918
-
'Generate high-quality videos from text prompts using leading AI providers. Supports Runway, Google Gemini, Luma, MiniMax, and Fal.ai multi-model generation with provider-specific durations, aspect ratios, resolutions, prompt optimization, and native audio controls.',
1934
+
'Generate high-quality videos from text prompts using leading AI providers. Supports Runway, Google Gemini, Luma, MiniMax, Higgsfield, and Fal.ai multi-model generation with provider-specific durations, aspect ratios, resolutions, prompt optimization, and native audio controls.',
0 commit comments