StarZero All shows Try the podcast clipper →

Mar 6, 2026 · 40m · startup-ideas

Using AI Tools (Manus, FreePik etc) to Create Viral IG Reels

Greg Isenberg · 12m spoken
0:00 / 0:00
▶ Watch on YouTube →

gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions

Host Greg Isenberg interviews viral creator Kova to reveal her end-to-end AI video creation pipeline, covering autonomous style reverse engineering with Manus AI, generative visual editing in Freepik, and automated storyboarding using Obsidian and Claude Code.

How this conversation actually went

Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Greg holds 36.2% of the talking time here. How this is scored →

Greg as informed peer 3.8 Guest teaching 5.4 Guest disagreement 1.2 Greg pushing back 1.2
05100:0015:0030:003:00–9:20 · Greg as informed peer 4/10 Demonstrating Autonomous Video Breakdown with Manus AI Greg suggests alternative agent workflows like direct file uploads and mentions OpenClaw beta releases. Kova explains the technical edge of autonomous agents like Manus over standard LLMs when breaking down video frames granularly.9:20–20:24 · Greg as informed peer 3/10 Analyzing the Style Guide and Five-Act Replication Framework Greg reads through the Manus AI replication breakdown while Kova educates him on aesthetic terminology like kawaii and her hero's journey build structure. Kova dismisses generic post-production checklists as unhelpful late-stage advice.20:25–28:54 · Greg as informed peer 4/10 Enhancing Static Backgrounds with Freepik and Masking Greg questions why a creator would bother altering a set that already looks decent, prompting Kova to reveal dramatic background transformations. Greg connects visual set depth directly to audience retention metrics and platform algorithmic distribution.28:54–33:44 · Greg as informed peer 3/10 Generating AI Video Transitions for Hook Retention Kova teaches Greg how generative video models parse tokens, warning against negative prompts like 'does not move' because models latch onto the raw keyword. Greg clarifies how prompt editors optimize text before stitching short AI clips together.33:45–39:20 · Greg as informed peer 5/10 Creator Differentiation and Full AI Tool Stack Overview Kova demonstrates using Claude Code and Cursor to programmatically convert scripts into storyboard tables inside Obsidian. Greg synthesizes the overall architecture, explaining how Manus style guides feed into Obsidian vaults for automated agent execution.3:00–9:20 · Guest teaching 4/10 Demonstrating Autonomous Video Breakdown with Manus AI Greg suggests alternative agent workflows like direct file uploads and mentions OpenClaw beta releases. Kova explains the technical edge of autonomous agents like Manus over standard LLMs when breaking down video frames granularly.9:20–20:24 · Guest teaching 6/10 Analyzing the Style Guide and Five-Act Replication Framework Greg reads through the Manus AI replication breakdown while Kova educates him on aesthetic terminology like kawaii and her hero's journey build structure. Kova dismisses generic post-production checklists as unhelpful late-stage advice.20:25–28:54 · Guest teaching 6/10 Enhancing Static Backgrounds with Freepik and Masking Greg questions why a creator would bother altering a set that already looks decent, prompting Kova to reveal dramatic background transformations. Greg connects visual set depth directly to audience retention metrics and platform algorithmic distribution.28:54–33:44 · Guest teaching 6/10 Generating AI Video Transitions for Hook Retention Kova teaches Greg how generative video models parse tokens, warning against negative prompts like 'does not move' because models latch onto the raw keyword. Greg clarifies how prompt editors optimize text before stitching short AI clips together.33:45–39:20 · Guest teaching 5/10 Creator Differentiation and Full AI Tool Stack Overview Kova demonstrates using Claude Code and Cursor to programmatically convert scripts into storyboard tables inside Obsidian. Greg synthesizes the overall architecture, explaining how Manus style guides feed into Obsidian vaults for automated agent execution.3:00–9:20 · Guest disagreement 1/10 Demonstrating Autonomous Video Breakdown with Manus AI Greg suggests alternative agent workflows like direct file uploads and mentions OpenClaw beta releases. Kova explains the technical edge of autonomous agents like Manus over standard LLMs when breaking down video frames granularly.9:20–20:24 · Guest disagreement 2/10 Analyzing the Style Guide and Five-Act Replication Framework Greg reads through the Manus AI replication breakdown while Kova educates him on aesthetic terminology like kawaii and her hero's journey build structure. Kova dismisses generic post-production checklists as unhelpful late-stage advice.20:25–28:54 · Guest disagreement 1/10 Enhancing Static Backgrounds with Freepik and Masking Greg questions why a creator would bother altering a set that already looks decent, prompting Kova to reveal dramatic background transformations. Greg connects visual set depth directly to audience retention metrics and platform algorithmic distribution.28:54–33:44 · Guest disagreement 1/10 Generating AI Video Transitions for Hook Retention Kova teaches Greg how generative video models parse tokens, warning against negative prompts like 'does not move' because models latch onto the raw keyword. Greg clarifies how prompt editors optimize text before stitching short AI clips together.33:45–39:20 · Guest disagreement 1/10 Creator Differentiation and Full AI Tool Stack Overview Kova demonstrates using Claude Code and Cursor to programmatically convert scripts into storyboard tables inside Obsidian. Greg synthesizes the overall architecture, explaining how Manus style guides feed into Obsidian vaults for automated agent execution.3:00–9:20 · Greg pushing back 1/10 Demonstrating Autonomous Video Breakdown with Manus AI Greg suggests alternative agent workflows like direct file uploads and mentions OpenClaw beta releases. Kova explains the technical edge of autonomous agents like Manus over standard LLMs when breaking down video frames granularly.9:20–20:24 · Greg pushing back 1/10 Analyzing the Style Guide and Five-Act Replication Framework Greg reads through the Manus AI replication breakdown while Kova educates him on aesthetic terminology like kawaii and her hero's journey build structure. Kova dismisses generic post-production checklists as unhelpful late-stage advice.20:25–28:54 · Greg pushing back 2/10 Enhancing Static Backgrounds with Freepik and Masking Greg questions why a creator would bother altering a set that already looks decent, prompting Kova to reveal dramatic background transformations. Greg connects visual set depth directly to audience retention metrics and platform algorithmic distribution.28:54–33:44 · Greg pushing back 1/10 Generating AI Video Transitions for Hook Retention Kova teaches Greg how generative video models parse tokens, warning against negative prompts like 'does not move' because models latch onto the raw keyword. Greg clarifies how prompt editors optimize text before stitching short AI clips together.33:45–39:20 · Greg pushing back 1/10 Creator Differentiation and Full AI Tool Stack Overview Kova demonstrates using Claude Code and Cursor to programmatically convert scripts into storyboard tables inside Obsidian. Greg synthesizes the overall architecture, explaining how Manus style guides feed into Obsidian vaults for automated agent execution.

speaking balance: gold is Greg, purple is the guest (3 minute bins)

0:00 · Greg 73.6% · guest 26.4%0:00 · Greg 73.6% · guest 26.4%3:00 · Greg 42.4% · guest 57.6%3:00 · Greg 42.4% · guest 57.6%6:00 · Greg 48.9% · guest 51.1%6:00 · Greg 48.9% · guest 51.1%9:00 · Greg 56.3% · guest 43.7%9:00 · Greg 56.3% · guest 43.7%12:00 · Greg 62.8% · guest 37.2%12:00 · Greg 62.8% · guest 37.2%15:00 · Greg 54% · guest 46%15:00 · Greg 54% · guest 46%18:00 · Greg 47.7% · guest 52.3%18:00 · Greg 47.7% · guest 52.3%21:00 · Greg 13.9% · guest 86.1%21:00 · Greg 13.9% · guest 86.1%24:00 · Greg 23.6% · guest 76.4%24:00 · Greg 23.6% · guest 76.4%27:00 · Greg 5% · guest 95%27:00 · Greg 5% · guest 95%30:00 · Greg 10.6% · guest 89.4%30:00 · Greg 10.6% · guest 89.4%33:00 · Greg 14.5% · guest 85.5%33:00 · Greg 14.5% · guest 85.5%36:00 · Greg 11.2% · guest 88.8%36:00 · Greg 11.2% · guest 88.8%39:00 · Greg 76% · guest 24%39:00 · Greg 76% · guest 24%
Sharpest disagreement ▶ 18:48 Rejecting AI-generated post checklists

Kova openly rejects the output from Manus regarding post-upload checklists, arguing creators should completely disregard it because foundational work must happen beforehand.

 Hardest push from Greg ▶ 23:53 Questioning necessity of background tweaks

Greg challenges the premise of spending time on AI inpainting by asking why anyone would bother when the original shot already looks completely fine.

 Biggest teaching moment ▶ 31:09 Technical mechanics of negative video prompts

Kova explains the underlying mechanics of diffusion models to Greg, showing why negative constraint words cause the model to generate the exact opposite of what was intended.

 Greg holds their own ▶ 38:50 Synthesizing the automated creator stack

Greg demonstrates strong domain fluency by synthesizing the entire end-to-end pipeline, articulating how Manus video extraction integrates into Obsidian markdown vaults and Cursor AI agents.
the scores for every segment, with the reasoning behind each
ChapterTopicGreg as informed peerGuest teachingGuest disagreementGreg pushing backWhy
Demonstrating Autonomous Video Breakdown with Manus AI 4411 Greg suggests alternative agent workflows like direct file uploads and mentions OpenClaw beta releases. Kova explains the technical edge of autonomous agents like Manus over standard LLMs when breaking down video frames granularly.
Analyzing the Style Guide and Five-Act Replication Framework 3621 Greg reads through the Manus AI replication breakdown while Kova educates him on aesthetic terminology like kawaii and her hero's journey build structure. Kova dismisses generic post-production checklists as unhelpful late-stage advice.
Enhancing Static Backgrounds with Freepik and Masking 4612 Greg questions why a creator would bother altering a set that already looks decent, prompting Kova to reveal dramatic background transformations. Greg connects visual set depth directly to audience retention metrics and platform algorithmic distribution.
Generating AI Video Transitions for Hook Retention 3611 Kova teaches Greg how generative video models parse tokens, warning against negative prompts like 'does not move' because models latch onto the raw keyword. Greg clarifies how prompt editors optimize text before stitching short AI clips together.
Creator Differentiation and Full AI Tool Stack Overview 5511 Kova demonstrates using Claude Code and Cursor to programmatically convert scripts into storyboard tables inside Obsidian. Greg synthesizes the overall architecture, explaining how Manus style guides feed into Obsidian vaults for automated agent execution.

Statements from this episode (0)

Nothing in this episode matches those filters. clear them

Made with StarZero

Turn any episode into a week of clips.

This entire site, over 300 episodes transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.

Try the podcast clipper Learn more