Skip to content

Add API media generation support across Pixelle workflows#120

Open
hit-cxf wants to merge 7 commits intoAIDC-AI:mainfrom
hit-cxf:feature/api-media-support
Open

Add API media generation support across Pixelle workflows#120
hit-cxf wants to merge 7 commits intoAIDC-AI:mainfrom
hit-cxf:feature/api-media-support

Conversation

@hit-cxf
Copy link
Copy Markdown

@hit-cxf hit-cxf commented Apr 30, 2026

Summary

This PR adds direct API-based media generation support to Pixelle-Video, enabling image, video, and VLM capabilities without requiring ComfyUI/RunningHub for supported workflows.

What Changed

  • Added API media adapters for image/video generation providers:
    • Image: DashScope Wan, Seedream, OpenAI image models
    • Video: DashScope Wan/HappyHorse, Kling, Seedance
    • VLM/LLM clients for asset analysis and prompt support
  • Integrated API media generation into major WebUI workflows:
    • Quick Create / Standard
    • Image-to-Video
    • Action Transfer
    • Digital Human
    • Asset-Based custom media workflow
  • Added API VLM-based asset analysis for the Asset-Based workflow, reducing dependency on ComfyUI for material evaluation.
  • Added dynamic video template support in Standard mode.
  • Improved video generation robustness:
    • Retry handling for transient network/download failures
    • Prompt safety retry for content-inspection failures by neutralizing prompts with LLM
    • Duration alignment with narration audio where applicable
    • Better continuity for Asset-Based video generation using previous segment tail frames
  • Added history persistence support for WebUI workflows so successful generations are recorded consistently.
  • Added API provider configuration examples in config.example.yaml.

Notes

  • Existing self-hosted/RunningHub workflows are preserved.
  • API workflows are exposed as additional provider-backed options.
  • Playwright remains required for HTML rendering; browser binaries still need installation via playwright install chromium.

Verification

  • Verified Standard quick-create image generation.
  • Verified Standard dynamic video flow wiring.
  • Verified Action Transfer API workflow.
  • Verified history persistence for multiple WebUI workflows.
  • Ran Python compile checks on modified modules during development.

@CLAassistant
Copy link
Copy Markdown

CLA assistant check
Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
You have signed the CLA already but the status is still pending? Let us recheck it.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants