Generates branded UGC videosCapturing the specified workflows and formats.
npx skills add https://github.com/TheMattBerman/scrollclaw --skill scrollclaw
Please help me install this Agent Skill. Skill: scrollclaw Source: https://github.com/TheMattBerman/scrollclaw/tree/main Command: npx skills add https://github.com/TheMattBerman/scrollclaw --skill scrollclaw
This Skill turns brand and audience inputs into AI-generated UGC videos that look like a real person recording on a phone, helping you avoid polished “AI-ad” output that viewers skip.
Use ScrollClaw to make a complete UGC video pipeline by providing your brand name and website URL, then asking for a UGC video for your target audience in the desired format.
High-intent search queries and answers about installing and using this skill.
To generate AI UGC videos with phone-photo realism, you need a pipeline that enforces handheld energy and first-frame control to prevent polished, skipable ad outputs. This approach orchestrates persona research, creator identity locking, and native captioning.
The best way to run a multi-step UGC campaign is to use a persistent pipeline that writes reusable brand and creator context. This allows you to orchestrate multiple content formats like talking head, hook+demo, and podcast clips while appending logs for faster iteration.
Yes, you need API keys for fal.ai or Replicate to handle the AI video generation and clip assembly. The pipeline also supports optional OpenRouter and ElevenLabs API keys for advanced routing and audio capabilities within the UGC workflow.
Yes, AI UGC video generation supports visual transformation, wall of text, talking head, hook face plus demo, podcast clip, and hybrid transformation formats. This format coverage ensures alignment with real UGC behaviors across different campaign workflows.
AI generated videos look too polished when they lack first-frame realism and handheld energy. An anti-polish creative system fixes this by locking creator identity, using real customer language, and applying UGC-native captioning as the final assembly step.
mattwoodcoGenerate videos from text prompts or images via Replicate and fal.ai.
TheMattBermanScore UGC candidate videos on seven virality criteria with a 70+ go/no-go gate.
different-aiGenerate podcast audio from text files using the ElevenLabs API.
kk20300113-pngAutomate raw video editing with FFmpeg cuts, scene detection, and platform reframing.
