What problem does it solve? Creating a two-host podcast-style video normally requires scripting, voice recording, filming, and editing; this Skill automates the entire pipeline from a single URL or topic into a finished 1-minute dialogue video. ## Core Features & Use Cases - URL and Topic Modes: Accepts a website, GitHub repo, or article URL (scraped for factual anchors) or a free-form brief with custom hosts, angles, and personas. - Four-Act Scripted Dialogue: Writes a Hook → Deep-dive → Pivot → Verdict script with jokes, interruptions, and a skeptic-flip moment, then renders four sequential 15-second Kling v3-omni acts with native two-voice audio. - Voice and Likeness Controls: Supports Kling preset voices, optional user voice cloning for Host A, custom host/background images, and archetype portraits instead of deepfakes for named public figures. - Use Case: Ask for a podcast review of a product launch page and receive a concatenated video URL of two hosts debating its features in about 8-18 minutes. ## Quick Start Ask the assistant to make a podcast about a specific URL or topic, for example: make a podcast reviewing https://pika.art with a skeptical second host.