richard900Q: Question about automation part 2
Since I ran out of space in part 1...
Any API integration with Kling or Minimax to create a video from a submitted image. So, for example, the workflow would generate an image (maybe using ChatGPT-4o), analyze it, and generate an image to video prompt. Then, it would be sent for approval. As part of the process, the approver would have the option to tweak the prompt and even swap out the image. Once approved, the workflow would receive the approval content and use Kling to generate a video based on the prompt and image.
Thanks in advance.
SeanP_AgenticFlowAI
Apr 24, 2025A: Hey Richard, you can absolutely wire this up today:
1. Kling/Minimax via Replicate
Connect your own Replicate key in the Replicate Connection (on the left sidebar)
(See it in action here: https://youtu.be/YGtIH1aegXw)
2. Generate & Analyze Image
– Use your LLM node (e.g. GPT-4o) to craft the image-to-video prompt
– Call Replicate to generate the initial image/video frames
3. Approval Step
– Drop in an email node (Gmail MCP: https://agenticflow.ai/mcp/gmail) or a simple form node to send the prompt/image to your approver
– Pause the workflow until you get their feedback (tweak prompt, swap image, etc.)
4. Final Video Generation
– Feed the approved prompt/image back into your Kling model via the same Replicate call
– Grab the resulting video and push it wherever you need
If you’d like a native “approval” node instead of the email workaround, please cast your vote here so we can bump it up our roadmap: https://agenticflow.featurebase.app/
Also this template is good too https://agenticflow.ai/app/explore/workflow/40b8c32a-cdd4-45fd-b369-5864915f2ac4 . Pixel ML offer Create Video from topic node which generate video with keyvisual from stock image / AI generated image