What You’ll Learn

James has 750 YouTube videos and, at the time of this session, exactly one public short. This is how he built a shorts pipeline he owns outright — no per-clip credits, no third-party service holding the output.

The Problem With Clip Tools

He’s clear that tools like Opus Clip do the cutting well. Three things pushed him elsewhere.

Reframing. Recording in landscape means something has to crop to vertical. Half the time he was out of frame or the screen content was cut, so the clips needed manual repair anyway.

Cost at volume. Fine for three or four videos a month. With a real library the credit maths turns into hundreds of dollars.

Distribution split. Clips live in one tool, the rest of your social lives in another. Now you’re posting from two places by hand.

Step 1 — Record Both Formats at Once

This is the highest-leverage change and it costs nothing.

Most streaming and recording software — Ecamm, OBS, StreamYard, Riverside — can output two formats simultaneously. James now records every session as landscape and vertical at the same time, each composed properly for its own shape.

💡 In Plain English: Instead of cropping a wide video down later and hoping, you record a vertical version designed as vertical. The reframing problem disappears because there’s nothing to reframe.

✓ Check Your Work: Open your recording software and look for a second output or “vertical” composition. If it’s there, turn it on before your next session — this only helps going forward.

Step 2 — Two Free Tools, Installed Locally

Both are open source and free, and Claude will install them for you if you ask.

Remotion builds video from code, which makes it very good at putting captions, titles, branding and overlays on top of existing footage. That replaces a round-trip through a separate editor.

HyperFrames — from HeyGen — composes video by writing HTML, CSS and JavaScript, which means an agent can drive it directly.

Step 3 — Know Which Tool Does What

James worked this split out explicitly, and it’s worth stealing:

• Face and voice talking → an avatar tool like HeyGen
• Body layers, screens, diagrams, captions, actual teaching → HyperFrames and Remotion
• Title cards and animated captions over a talking head → Remotion
• Something generated from nothing → a generative video tool

Step 4 — Templates, Not Clips

The framing that makes this scale:

“Think of these as fill-in-the-blank templates. Build the look once, then pour in new words forever.”

Because the campus already holds the brand — fonts, colours, voice — the template gets built on-brand once. The formats he identified: kinetic hook shorts, tips lists, stat reveals, face-lit concept explainers, quote and social proof cards, and event announcements.

Several of those need zero footage. That’s the difference between an automated system and a video editor.

Step 5 — Distribute From One Place

The renders go to the same social scheduler as everything else, already connected to the vertical platforms. One central place to hand content to, rather than one pipeline per tool.

Two Honest Caveats

⚠️ This pipeline produced unusable clips on its first run. The mechanics worked and the cut boundaries were wrong. James covers that failure in a separate session, and it’s the necessary companion to this one — build the factory, but do not switch off the review gate.

⚠️ Recording dual-format only helps future sessions. Your back catalogue still needs reframing or a transcript-driven template that uses no original footage.

Key Takeaways

• Record vertical and horizontal simultaneously — most streaming software already supports it and it removes the reframing problem entirely.
• Per-clip credit pricing breaks down at library scale; local open-source rendering doesn’t.
• Use Remotion for captions, titles and overlays on existing footage, and HyperFrames when an agent needs to compose video directly.
• Build one on-brand template and pour new words into it, rather than editing each clip as a one-off.
• Route renders through the same scheduler as the rest of your social, so distribution stays in one place.

Your Next Step

Turn on the second output in your recording software today. Everything else in this pipeline is easier once you have footage that was born vertical.

The Campus AI OS is free and holds the brand context these templates read from: trainingsites.io/os