Lock the format and story job
Define the viewer, promise, duration, aspect ratio, emotional movement, and final action before the asset list exists.
For the video that needs more than a folder of generated clips
Move from one-page brief to a finished video through twelve visible stages.
Video combines story, images, narration, music, timing, captions, color, sound, and delivery. The pipeline makes each handoff inspectable, so the project can move forward without pretending one automated edit replaces the decisions.
One-time purchase through Whop. Training, resources, and downloads stay together.
AI Video$49
Starting statethe video that needs more than a folder of generated clips
- 01Lock the format and story jobDefine the viewer, promise, duration, aspect ratio, emotional movement, and final action before the asset list exists.
- 02Build the edit in layersCreate and log assets, record narration, assemble the first pass, and iterate on structure before polishing every transition.
- 03Finish for deliveryBalance audio, establish visual consistency, generate and review captions, render, inspect, and package the master for its destination.
Reviewable outputThis toolkit moves through brief and vibe, shot list, asset preparation, script and voiceover, first-pass edit, iteration, timing, audio mix, color, captions, delivery, and review.
The failure pattern
Output is not the same as a finished system.
AI video projects often begin by generating assets before the format, audience, duration, and narrative job are clear. The result is expensive drift: beautiful shots that do not cut together, narration with no room to breathe, a music bed that fights the voice, and a final render nobody has checked against the original promise.
The change
Make each handoff visible.
This toolkit moves through brief and vibe, shot list, asset preparation, script and voiceover, first-pass edit, iteration, timing, audio mix, color, captions, delivery, and review. It uses Kevin's actual stack: fal stills, Kokoro narration, ACE-Step music, Remotion, local Whisper, and FFmpeg.
Build the edit in layers
Create and log assets, record narration, assemble the first pass, and iterate on structure before polishing every transition.
Finish for delivery
Balance audio, establish visual consistency, generate and review captions, render, inspect, and package the master for its destination.
Inside the toolkit
Reference, guidance, and files you can actually inspect.
Twelve-stage video manual
The front and back halves of the pipeline, plus genre modules for book trailers, app demos, and planned vertical work.
Skills, commands, and templates
A brief, shot list, edit notes, color reference, release checklist, and stage-specific workflow files.
Sleep Mode worked example
The real brief, production log, source code, narration, images, music, and render decisions behind a completed book trailer.
Source-backed
The receipt
A real trailer sits underneath the entire pipeline.
The rebuilt written toolkit is grounded in the Sleep Mode book trailer, not an invented client project. Its source folder preserves the brief, shot generation, Kokoro narration, ACE-Step music bed, Remotion code, and render history used by the toolkit. The older standalone course-video modules are not included while their truthful rebuild remains pending.
Explore Kevin's production workStrong fit
This is for you if...
- You are producing a trailer, demonstration, promo, or structured social video.
- You want a tool-aware workflow without making the course about one editing interface.
- You need the project files and review gates that connect generation to delivery.
Wrong fit
Skip it if...
- You want random generated clips with no brief, edit, or review process.
- You expect one tool to replace creative direction, sound decisions, and final inspection.
Before checkout
Straight answers.
Which tools does Kevin use?
The worked pipeline uses fal for stills, Kokoro for local narration, ACE-Step for music, Remotion for the edit, local Whisper for captions, and FFmpeg for media checks and finishing.
Do I need traditional editing software?
No specific editor is required. The course teaches the shape and checkpoints of the decisions, with a code-driven worked example.
Does it teach vertical video?
The vertical module is clearly presented as a method and plan because Kevin had not yet shipped that format when the core product was audited.
Can I use my own footage?
Yes. The pipeline supports generated stills, screen recordings, stock, camera footage, or mixed sources. The balance and consistency work changes with the source material.
Are the standalone course videos included now?
No. The current delivery includes the written manual, browser reading pages, Resources, Downloads, templates, commands, Skills, and the Sleep Mode worked example. The standalone course-video rebuild remains pending and is excluded from the current delivery.
The boundary
The current delivery is the written production toolkit, browser reading pages, Resources, Downloads, templates, commands, Skills, and completed Sleep Mode worked example. Standalone course videos remain pending and are not included. The toolkit does not promise views, conversion, virality, or a finished video without creative and technical decisions.
AI Video Pipeline Toolkit