How to Edit VIRAL Podcast Reels ONLY Using AI! (Full Breakdown)

Overview

The video is a full, step-by-step A-to-Z tutorial for creating viral Instagram/TikTok-style podcast reels using only AI, even for beginners. It presents a five-phase workflow using a combination of AI tools (mostly Claude, plus 11Labs, Submagic, and Hicksfield / “Cance 2.0” for animation and design).

---

Five-phase AI workflow (core tutorial)

Phase 1 — Clipping: extract viral short-form moments

Goal: Turn a long-form podcast video into 5 “premium” short clips automatically.

Inputs used:

1. 11Labs (speech-to-text) generates a transcript from the podcast YouTube link.
2. Generate timestamps by creating subtitles and exporting as JSON (JSON includes “what is being said” + “when”).
3. Upload the JSON to Claude.

Claude capability: “short form clipper skill”

• Extracts five high-quality clips from the transcript.
• The creator reviews and selects favorites (example: clip three), noting the best clip typically has:
  • a relatable hook,
  • an unexpected reframe,
  • storytelling structure (example + outcome).
• Claude then trims each clip from the original video using aligned timestamps.

Key claim: avoids manually watching the entire podcast to find hooks; Claude handles clip extraction.

---

Phase 2 — Captions: auto-generate subtitle styling

After clipping, the tutorial generates captions using Submagic (alternatives mentioned: Dscript, Opus).

Steps in Submagic:

• Create project → upload the clipped video
• Choose a caption style (example used: clean/minimal, “Kelly 3” style)
• Customize typography:
  • font: Poppins
  • light weight
  • size scaled to ~30
  • position: lower thirds
• Export/download the subtitled clip

Key goal: fast captioning + a clean, readable style that matches minimal animation edits.

---

Phase 3 — Planning the edit for retention/virality

The creator uses Claude to plan what should appear on screen to maximize:

• viewer retention
• optimal virality

Claude runs a “podcast clip planner / editor brain skill” (trained from prior examples/courses).

Planning includes mapping spoken moments to visuals, such as:

• A-roll (talking head-style framing) for connection
• graphics/design scenes (e.g., dumbbell, TV, forum-reading)
• time jumps (planned visually, while leaving some sections as A-roll for rest)
• conclusive/ending visuals for payoff

Example structure shown:

• hook → dumbbell scene → TV scene → time-jump/return beats → fossil/orbit/identity visual metaphors → concluding end screen.

Key claim: planning is treated like a professional editor workflow, using “content psychology” and scene logic.

---

Phase 4 — Design: turn the plan into vertical motion-ready artwork

Claude generates scene designs from the plan using a “podcast designer skill.”

Design direction:

• color theme: white + orange
• output: designs for every planned scene in Hicksfield
• generation detail: “GPT image 2 at 2K resolution”

Training/inspiration approach:

• Claude is fed lots of reference examples (the creator mentions gathering hundreds of inspirations from Pinterest) so it learns:
  • layout patterns
  • color psychology
  • vertical reel design style

Creator reviews produced designs, including:

• dumbbell + text animation frame
• TV scene
• fossilized dumbbell design
• orbit scene
• additional symbol scenes (e.g., “scarf”)
• final conclusive scene

Key output: premium-looking, consistent vertical design assets that match the planned visuals.

---

Phase 5 — Animation: convert designs into motion

Claude animates the designs using Hicksfield / “Cance 2.0.”

Prompting details (as described):

• use podcast animation skill
• animate into high-quality short-form minimal animations
• use “Cance 2.0 at 1080p”
• generate with no music, only sound effects

Animation logic (example timing/style):

• object slides in (realistic stone effect)
• text pop-up / blur fade-in, word-by-word
• edge graphics animate in
• motions are quick, smooth, seamless, avoiding “rigid/bouncy” movement

Workflow timing/integration:

• Claude studies animation prompt examples and converts behavior into a reusable “podcast animator skill.”
• Animations are placed into Premiere Pro at the exact planned timeline locations.
• Manual finishing (post-production):
  • adjustment layer with zoom-out
  • whooshes/risers for the first 5 seconds (hook)
  • cut a slow section and track the subject for a better effect
  • add a song to bring it together

---

Cost + time breakdown (pricing/efficiency data)

Credits usage (Hicksfield)

• “8 images at 7 credits each”
• “8 animations at 55 credits each”
• total: just under 500 credits

Estimated cost

• top-up ~500 credits for about $26
• total reel cost: ~$25

Time (end-to-end)

• clipping: 12 min
• captions: 5 min
• planning: 8 min
• design: 10 min
• animations: ~16 min
• final touches: ~15 min
• Total: just over an hour, but much of that is background AI processing
• “Real work” stated: ~15–20 minutes, plus AI rendering time.

---

Key takeaway / promised result

By following the workflow, the creator claims you can generate a premium, on-brand viral podcast reel quickly and consistently:

• automated clipping
• styled captions
• virality-focused scene planning
• AI design generation
• AI animation generation
• light manual finishing

---

Main speakers / sources (as presented)

• Main speaker/host: the tutorial creator (uses Chris Williamson as the example podcast source)
• Podcast source referenced: Chris Williamson

AI/data sources/tools mentioned:

• Claude (core orchestrator for clipping, planning, design logic, animation logic)
• 11Labs (speech-to-text + transcript + JSON export with timestamps)
• Submagic (captions)
• Hicksfield / “Cance 2.0” (scene design + animation rendering)
• Premiere Pro (timeline assembly + final edits)