# The AI stack behind my viral content (10M+ views) Source: https://www.youtube.com/watch?v=0b8qQx3FaLE Recap page: https://rapidrecap.app/video/0b8qQx3FaLE Generated: 2026-03-06T20:03:15.928+00:00 --- ## Quick Overview Creator Cova shares her detailed, step-by-step workflow for creating highly engaging, viral short-form video content, emphasizing the use of AI tools like Midjourney (for image generation) and Seodance 1.5 Pro (for video generation), combined with manual editing in Adobe Premiere Pro and robust documentation in Obsidian for scalability. **Key Points:** - Cova successfully replicated a viral video style using AI video generation (Seodance 1.5 Pro) by providing it with a detailed prompt based on aesthetic keywords, script transcription, and visual breakdown. - The successful replication involved giving the AI a detailed prompt specifying visual elements like "orangutan plush sitting on the windowsill" and enforcing a static camera shot. - For asset creation, Cova uses FreePick for stock images/videos and Adobe Premiere Pro for complex visual effects like masking and compositing, although she notes simpler effects can be done in CapCut. - Cova relies heavily on structured documentation within Obsidian, using templates for self-storyboarding and style guides (like the 'Covacut style guide') to ensure consistency across all projects. - The video's structure follows a classic 5-act build-log structure (Hook, Conflict/Stakes, Build, Problem/Resolution, Tease/CTA) with a target runtime of 60-75 seconds. - She emphasizes the importance of creating a strong 'bit' (one-sentence concept hook) that is emotionally resonant and explainable in under 10 seconds. - The process involves using AI for initial generation (like adding the plush or tulips) but relies on manual editing for precise masking and layering, such as isolating elements from the original footage. ![Screenshot at 00:00: The host introduces creator Cova, who is known for viral short-form videos, setting the stage for a detailed breakdown of her AI-enhanced content creation workflow.](https://ss.rapidrecap.app/screens/0b8qQx3FaLE/00-00-00.jpg) **Context:** This video is an interview between Greg Isenberg and creator Cova, who is known for creating viral short-form content (Reels/TikToks) that blend hacker culture aesthetics with 'kawaii nostalgia' and 'cinematic bedroom aesthetics.' Cova details the technical and conceptual framework she uses to consistently produce high-quality, complex videos, often leveraging AI tools for initial asset creation and visual inspiration, before refining the output using professional software like Adobe Premiere Pro and organizing her entire workflow using Obsidian. ## Detailed Analysis Cova details her comprehensive workflow for creating viral short-form videos, which relies on a synergistic blend of AI tools and meticulous manual organization. She breaks down the process into eight main steps, starting with defining a strong 'Bit' (concept hook) that is emotionally resonant and concise (07:08). For visual style, she relies on aesthetic keywords like 'dark academia maker,' 'kawaii nostalgia,' and 'cinematic bedroom aesthetics' (09:47). In the video generation phase, she uses tools like Seodance 1.5 Pro, providing highly specific prompts (e.g., 'Camera is static. On the table is a picture of a kid. The kid is waving their arms gleefully while the photo doesn't move.') to guide the AI (30:24). For B-roll and VFX, she uses Adobe Premiere Pro for detailed masking and compositing, especially for the technically demanding 'Physical is Back' shot (16:56), although simpler tasks can be done in CapCut (27:09). Her organizational structure is critical: she uses custom templates in Obsidian for storyboarding, referencing her 'Covacut style guide' for typography and sound design consistency (36:32, 37:44). The final video structure adheres strictly to a 5-act structure (Hook, Conflict, Build, Problem/Resolution, Tease/CTA) targeting 60-75 seconds (12:58). ### Video Aesthetic & Concept - Video blends hacker culture with kawaii nostalgia and cinematic bedroom aesthetics - Key aesthetics include moody low-key RGB lighting, specific color palettes (deep black, neon orange, warm amber), and distinct typography styles (chunky rainbow gradient font for titles) - The core concept must be explainable in under 10 seconds. ### Workflow Tools Used - AI video generation via Seodance 1.5 Pro (preferred over other models for her style) - Image generation via Freepik or AI image tools - Editing primarily in Adobe Premiere Pro for complex effects (masking, compositing) - Organization and documentation in Obsidian using custom templates (Self Storyboard Template, Covacut style guide). ### Video Structure (5-Act Build-Log) - Hook (0-10s, show finished product first) - Conflict/Stakes (10-22s, introduce personal challenge) - Build (22-43s, show steps) - Problem & Resolution (43-54s, emotional payoff) - Tease & CTA (54-65s, low-pressure call to action) - Total runtime target: 60-75 seconds. ### Technical Execution (B-Roll & VFX) - B-roll requires three types: hardware macro close-ups, screen recordings (Photoshop/Figma/VS Code), and product reveal shots - VFX 'Physical is Back' shot requires green screen footage, sourcing retro tech PNG cutouts, layering in post-production (After Effects/DaVinci Resolve), and adding subtle motion blur. ### Audio & Music - Use a lo-fi, slightly nostalgic background track (70-90 BPM) with a warm, crunchy texture (chiptune/lo-fi hip-hop) - Music kept at -18 to -20 dB under the voice - Subtle UI/tech sound effects are added on title cards and section reveals. ![Screenshot at 00:01: Demonstration of AI video generation capabilities, showing the transformation of a scene using AI prompts.](https://ss.rapidrecap.app/screens/0b8qQx3FaLE/00-00-01.jpg) ![Screenshot at 00:25: Cova displays her Instagram profile, highlighting several high-view-count videos that exemplify her distinct aesthetic.](https://ss.rapidrecap.app/screens/0b8qQx3FaLE/00-00-25.jpg) ![Screenshot at 03:17: Cova demonstrates using the Prompt Editor in the video generator, showing how specific, detailed descriptions help guide the AI output.](https://ss.rapidrecap.app/screens/0b8qQx3FaLE/00-03-17.jpg) ![Screenshot at 04:45: Cova navigates her Obsidian vault, showing detailed project documentation including a 'Video Breakdown & Replication Plan' that outlines every shot's concept and technical requirements.](https://ss.rapidrecap.app/screens/0b8qQx3FaLE/00-04-45.jpg)