Keep source media in your project
Register footage by path. The pipeline stores project records locally and does not upload your recordings to this website.
project.db · media stays yours
LOCAL VIDEO WORKFLOW · HUMAN APPROVAL
From raw recordings
to a review queue.
An advanced online playbook for teams processing their own long-form video. Follow the setup steps and copyable Python commands to transcribe, rank candidate moments, suggest portrait crops, render batches, and review each decision.
Advanced online playbook + copyable code · access on this website after payment · $10,000 USD
studio / review queueLOCAL
00:02:18
Why the first draft failed00:02:18 — 00:02:52 · 84 pts
00:06:42
Three costs teams forget00:06:42 — 00:07:13 · 77 pts
00:11:03
A better way to test it00:11:03 — 00:11:38 · 72 pts
3 candidates · 1 approved · 0 publishedreview is the gate ●
01Ranking helps sort the queue.
It never makes the editorial decision.
It never makes the editorial decision.
01 local project database
02 transcription + clip scoring
03 crop suggestion + render queue
04 manual channel publishing
A four-stage workflow
Repeat the operations.
Keep the judgement.
Designed for a local team workflow with an explicit approval step. Candidate scoring is a sorting aid, not a claim that any clip will perform.
Transcribe and rank moments
Run local Whisper transcription, then score bounded clip windows using clear, editable heuristics. Inspect every candidate in context.
transcript · candidate.json
Choose what deserves a cut
Review transcript, full context, caption text and crop. Approve specific candidates; unapproved clips are not rendered.
explicit human approval
Render a traceable batch
Use brand profiles, caption sidecars, run records and upload notes to prepare consistent exports for your normal channel workflow.
MP4 · SRT · audit log
Designed for the work after the demo
A codebase with
the handoff included.
The online materials explain environment setup, project operations, configuration, batch review, and what still requires a human decision.
Repeatable batch jobs
Record source, profile, candidates, status and errors in a local SQLite project.
Face-aware crop suggestion
Sample frames to suggest a portrait crop. If no face is found, use a centered crop and inspect it.
Brand profiles + output
Keep caption styles and output conventions consistent across a channel or client.
Human review stays required
Approve candidates and inspect exports before anyone posts or makes a factual claim.
Digital materials
Online playbook,
copyable code.
The complete playbook and code library open here after payment is confirmed. Sign in with your email and password to return to your purchases. Your account keeps the materials together; video processing runs on your own computer.
Production playbook
Step-by-step guidance for editorial selection, transcript review, captioning, batch production, rights checks, and release operations.
READ · APPLY · REVIEWCopyable Python code
Project setup, transcript processing, explainable candidate scoring, approval, crop suggestions, batch rendering, and run records.
PYTHON · SQLITE · FFMPEGSetup and operating steps
Environment checklist, brand configuration, command examples, troubleshooting notes, and a sample run using footage you provide.
NO SAMPLE USER FOOTAGECommand walkthrough
Analyze first.
Approve before rendering.
The advanced workflow stores jobs locally. Candidate rankings help organize review; they do not decide what deserves publication.
Initialize a project and register your own video:
python shortsstudio.py init ./studio-project
python shortsstudio.py add ./studio-project --video /path/to/your-video.mp4 --brand MainAnalyze, inspect, approve selected candidates, then render:
python shortsstudio.py analyze ./studio-project --job 1 --model small
python shortsstudio.py candidates ./studio-project --job 1
python shortsstudio.py approve ./studio-project --job 1 --clips 1,3
python shortsstudio.py render ./studio-project --job 1Requires Python 3.11+, FFmpeg, and optional local model packages. Inspect the transcript, crop, captions, and every export yourself.
Studio license
Built for a working production team.
The one-time digital product provides the online playbook, full copyable code library, setup instructions, and a commercial-use license for one organization’s internal workflow. The complete library opens on this website after payment is confirmed; sign in with your email and password to return later. It does not include YouTube approval, cloud hosting, model/API fees, media rights, or guaranteed business results.
Review system requirements and the digital-product terms before purchasing. Secure payment and immediate library access are available on this website.
ONE-TIME$10,000Buy digital access ↗
One-time purchase. Review the product and delivery terms on our checkout page.
Limits and requirements
What it does—and what it leaves to you.
Does it post or scrape?
No. It works on local media and exports review drafts. It does not scrape other channels, automate engagement, or publish without your approval.
What do I need to run it?
Python 3.11+, FFmpeg, and optional model packages for local transcription and face detection. The setup guide lists disk, memory, and model download trade-offs.
Will this make videos go viral?
No software can promise that. Candidate scoring organizes review work; it does not know the audience or guarantee results.
Can I auto-upload to YouTube?
Uploads are outside this package. API projects and channels must satisfy YouTube's current verification and policy requirements before publishing integration can be considered.