Trevor Crane / Epic · built on open source

Epic Clip Machine

One long video in. Scroll-stopping shorts out, in English, Spanish and Portuguese.

9finished shorts
3×3clips × languages
$0monthly software (the AI steps bill per use)
1command per new video

The 6-step machine

One long video in. Finished English, Spanish and Portuguese shorts out, with captions, thumbnails and copy ready for review. Every step is a free open-source tool except the two AI steps.

  1. 01

    Fetch

    yt-dlp pulls the original video. The source file is never edited or overwritten.

    yt-dlp
  2. 02

    Transcribe

    faster-whisper writes a word-by-word transcript with exact timings (JSON + SRT).

    faster-whisper
  3. 03

    Pick

    An AI reads the transcript and picks the strongest 20–45s moments. Timecodes are checked against real words; invented ones get rejected.

    LLM + checker
  4. 04

    Cut + caption

    Each moment is cut to 9:16 from the original, with bold word-timed captions (2 lines, 42 characters max) and a hook title.

    ffmpeg + libass
  5. 05

    Translate + review

    Spanish (LatAm) and Brazilian Portuguese with the same cue timing, a locked glossary of your phrases, and an automatic check for line length and glossary terms.

    LLM + glossary
  6. 06

    Package + review

    Every clip in every language gets an MP4, SRT, thumbnail, title, caption, hashtags and alt text. A frame from every caption is checked. Nothing is posted until you say yes.

    ffmpeg + checks

First run: 3 clips, 3 languages

Source: Create Content Quickly (for Busy Business Owners). Tap a language to switch the video, captions and copy. These are drafts for your review. Nothing has been posted anywhere.

All clips:
CLIP 01

Content doesn't get made without a deadline

Universal pain point, relatable teen-homework analogy, one clear action (pull out your calendar).
On-screen hook
No deadline = no content
Title
Content doesn't get made without a deadline
Caption
Even my 15-year-old gets her homework done just in time. Step one to creating content quickly: set a deadline. Pull out your calendar and decide when it's going to be done.
Hashtags
#ContentCreation #BusinessOwner #Deadlines #Productivity #Entrepreneur #AuthorLife #TrevorCrane
Alt text
Trevor Crane talks to camera in his office, explaining that the first step to creating content quickly is setting a deadline on your calendar.
CLIP 02

One weekend, one piece of content, a course AND a book

A real story with a payoff: brainstorm day one, create day two, it became a course and then a book.
On-screen hook
1 weekend = a course + a book
Title
One weekend, one piece of content, a course AND a book
Caption
My friend Steve Napolitan came to visit, so I said: spend the night and let's create something for our clients. Day one we brainstormed. Day two we created. It went into a course and turned into this book.
Hashtags
#ContentCreation #Repurpose #OnlineCourse #WriteABook #BusinessOwner #Entrepreneur #TrevorCrane
Alt text
Trevor Crane tells the story of a friend's overnight visit that produced one piece of content, which became a course and then a book.
CLIP 03

Dare to suck: progress beats perfect

Strongest catchphrase in the video, emotional permission, sticky-note proof. Same moment as the hand-made proof clip, so it doubles as a parity test.
On-screen hook
Dare to suck
Title
Dare to suck: progress beats perfect
Caption
Step four: just do it. I wrote today's ideas on a sticky note. Keep it simple, go out there and dare to suck. You don't have to be perfect, you just want to make progress and get it done.
Hashtags
#DareToSuck #JustDoIt #ContentCreation #ProgressNotPerfection #BusinessOwner #Entrepreneur #TrevorCrane
Alt text
Trevor Crane holds up the sticky note he wrote his ideas on and tells viewers to dare to suck, because progress beats perfection.

Quality checks

Burn-in check: PASS. One frame was pulled at every caption cue of every finished short and compared with its SRT file.
  • 01-set-a-deadline en: 13 cues; srt==burned text: True; text on screen at every cue (min 2.2% of caption band changed): True -> PASS sheet: verify/01-set-a-deadline/en/sheet.jpg
  • 01-set-a-deadline es: 13 cues; srt==burned text: True; text on screen at every cue (min 1.1% of caption band changed): True -> PASS sheet: verify/01-set-a-deadline/es/sheet.jpg
  • 01-set-a-deadline pt-BR: 13 cues; srt==burned text: True; text on screen at every cue (min 1.5% of caption band changed): True -> PASS sheet: verify/01-set-a-deadline/pt-BR/sheet.jpg
  • 02-one-weekend-course-and-book en: 27 cues; srt==burned text: True; text on screen at every cue (min 2.2% of caption band changed): True -> PASS sheet: verify/02-one-weekend-course-and-book/en/sheet.jpg
  • 02-one-weekend-course-and-book es: 27 cues; srt==burned text: True; text on screen at every cue (min 2.7% of caption band changed): True -> PASS sheet: verify/02-one-weekend-course-and-book/es/sheet.jpg
  • 02-one-weekend-course-and-book pt-BR: 27 cues; srt==burned text: True; text on screen at every cue (min 3.0% of caption band changed): True -> PASS sheet: verify/02-one-weekend-course-and-book/pt-BR/sheet.jpg
  • 03-dare-to-suck en: 15 cues; srt==burned text: True; text on screen at every cue (min 1.0% of caption band changed): True -> PASS sheet: verify/03-dare-to-suck/en/sheet.jpg
  • 03-dare-to-suck es: 15 cues; srt==burned text: True; text on screen at every cue (min 2.0% of caption band changed): True -> PASS sheet: verify/03-dare-to-suck/es/sheet.jpg
  • 03-dare-to-suck pt-BR: 15 cues; srt==burned text: True; text on screen at every cue (min 1.2% of caption band changed): True -> PASS sheet: verify/03-dare-to-suck/pt-BR/sheet.jpg

Method: `ecm.py verify` pulled one frame per cue (cue midpoint) from each finished mp4 and laid the caption band

next to the expected .srt text (contact sheets: verify/<clip>/<lang>/sheet.jpg, combined: verify/<clip>-all.jpg).

I read every row of all 9 sheets.

Result: 165 / 165 cue frames show exactly the .srt text, same line breaks, no missing, extra or stale cues.

- 01-set-a-deadline: en 13/13, es 13/13, pt-BR 13/13

- 02-one-weekend-course-and-book: en 27/27, es 27/27, pt-BR 27/27

- 03-dare-to-suck: en 15/15, es 15/15, pt-BR 15/15

Accents render correctly (é, ñ, ã, ç, ¿, ¡). Source's own English captions/lower-thirds are trimmed out of frame

(runs/2k01Q2Vvr2A/overrides.json), so ES/PT shorts show no English caption text.

Note: in 03 cue 5 the source's own big orange "RIP IT" graphic is partly visible at the edge of the frame (it is part of the video picture, not our captions).

Translation review: same model. Grok Bot translated and then reviewed its own work. That is a self-check, not an independent Native Review. A second AI model or a native speaker should review before anything goes live.
  • SAME-MODEL REVIEW: the translator and the reviewer are the same model. This is NOT an independent Native Review. A second model family or a native speaker should review before anything ships.
  • Cue counts match English exactly (13/27/15); timing reused unchanged. Every cue fits 2 lines x 42 chars (checked by the translate step).
  • LOCKED glossary terms used: solo hazlo / só faz, atrévete a hacerlo mal / ouse ser ruim, ¡Lánzate! / Manda ver!, sacarlo adelante / fazer acontecer, post-it.
  • Clip 03 ES/PT reuse the Native-Review-passed proof phrasing (Nov 18, es 92-93, pt-BR 92) where the English matches; cue splits differ because the English cueing is new.
  • Clip 02: 'It is how we were gonna use it' rendered by meaning ('Así decidimos cómo usarlo' / 'Foi assim que decidimos usar') since a literal version is unclear. Cues 1-2 reorder the sentence for natural word order; each cue still covers the same spoken span.
  • Clip 01: 'one of the most important things is to set when is it going to be released' compressed to 'Esto es de lo más importante: definir cuándo se va a publicar' / 'Essa é uma das coisas mais importantes: definir quando vai ser lançado'.
  • Register: ES tú (LatAm), PT você casual with 'pra/pros/tá/tava'. Hashtag volumes NOT checked (no TikTok/Instagram login).

Licenses: can we bundle & resell it?

Short answer: yes. Everything is permissive except FFmpeg/x264 (GPL in this build). We can call FFmpeg freely and own the videos it makes. If we ship the FFmpeg binary inside a paid bundle, we also have to offer its source code. Checked Nov 18 2026: license field from the GitHub API for each repo, plus the FFmpeg LICENSE.md, x264 COPYING and Pillow LICENSE files and the Hugging Face model card.

ToolLicenseWhat it means for bundling / reselling
yt-dlp
1. Fetch (download)
Unlicense (public domain)Free to bundle, rebrand and resell. No notice required. Only download videos you own or have rights to (YouTube terms). source
faster-whisper
2. Transcribe
MITFree to bundle and resell. Keep the MIT copyright notice in the bundle. source
CTranslate2
2. Transcribe (engine)
MITFree to bundle and resell with the notice. source
Whisper model weights (Systran/faster-whisper-medium, from OpenAI Whisper)
2. Transcribe (model)
MITFree to ship and use commercially with the notice. source
FFmpeg (this box's build)
4/7. Cut + burn captions
LGPL-2.1+ by default; GPL-2.0+ when built with --enable-gpl. This box's build has --enable-gpl and libx264, so it is GPL.Calling ffmpeg as a separate program is fine, and the videos it makes are yours. If we ship the ffmpeg binary inside a paid bundle we must also offer its source code (GPL). Cleanest: the installer has the customer install ffmpeg, or we ship an LGPL build. The pipeline avoids GPL-only filters (uses gblur, not boxblur). source
x264 (H.264 encoder inside ffmpeg)
7. Burn / encode
GPL-2.0Same GPL rule as above if we ship it. x264 also sells a commercial license (UNVERIFIED: not checked today). H.264 patent licensing is a separate question (UNVERIFIED). source
libass
7. Caption renderer
ISCPermissive: bundle and resell with the notice. source
Pillow
Text fitting, verify sheets
MIT-CMU / HPND-style permissive (GitHub API says NOASSERTION; read from LICENSE file)Permissive: bundle and resell with the notice. source
PyAV
Installed with faster-whisper
BSD-3-ClausePermissive. Note its wheels bundle FFmpeg libraries (LGPL). source
NumPy
Audio buffer
BSD-3-Clause (UNVERIFIED today) UNVERIFIEDExpected permissive; not checked today. source
Montserrat font (caption font; static ExtraBold instance renamed 'Montserrat ECM')
Captions
SIL OFL 1.1Can ship inside paid software. Can't sell the font by itself. Our modified copy must stay under OFL. source
Anton font
Hook titles
SIL OFL 1.1Same as Montserrat. source
Wrangler (Cloudflare deploy CLI)
Deploy only
Apache-2.0 (workers-sdk repo)Not shipped in the bundle; used to publish this page. source
LLM for pick + translate (Grok Bot today; Claude or any model via ECM_LLM_CMD)
3. Pick, 5. Translate + review
Not open source. Paid per use.The one step with a running cost. Customers bring their own model key, or Epic covers usage.
Epic Clip Machine code (ecm.py, burn.sh, site)
The machine
Epic's own code. No license file yet: Trevor's call (private / proprietary bundle).Ours to brand and sell.

This is a plain-English summary, not legal advice.

Run it on a new video

On the box, one command. It stops and asks for the two AI answers (moment picks, translations) unless an AI command is set with ECM_LLM_CMD; re-running picks up where it left off.

cd /workspace/epic-clip-machine
python3 ecm.py run "https://www.youtube.com/watch?v=VIDEO_ID"
# optional: let a CLI model answer the AI steps automatically
ECM_LLM_CMD='claude -p' python3 ecm.py run "https://www.youtube.com/watch?v=VIDEO_ID"
# rebuild + redeploy this page
python3 site/build_site.py && ./deploy.sh
Copied