用 Higgsfield Genjutsu + Hedra 做换脸视频的 bot 工作流
用户分享用 Higgsfield Genjutsu 做全身换脸、Hedra 做口型同步的自动化流程，靠 bot 先出预览图再渲染，称预览步骤最省积分。
AI video generation is in my top 3 use cases for data center water and here's the workflow I've found most helpful. 1/ Tools: @higgsfield_ai Genjutsu for full-body swaps, Hedra for lip-sync, and @bot running all of it. 2/ My setup is a Chief of Staff bot that hands video jobs to a dedicated AI Video Desk bot. I signed in to Higgsfield once and haven't logged in since. 3/ The flow: Send the reel and say who goes in it The bot pulls the video and saves the original audio It builds a preview from real photos, never made-up faces I approve the preview before anything renders It runs the Genjutsu swap It puts the audio back on, checks the lyrics, and sends me the final 4/ The preview step saves the most credits. Most bad renders come from a bad starting image. 5/ Want the same setup? Here's my bot template: https://x.ai/bot/EXmZpPdhI7H4LoCjwwhyk