The need
He ran Claude on video editing repeatedly and the cuts still weren't good enough to be happy with, and only found an acceptable path by discovering a third-party studio layer himself.
Original post — verbatim Thai
ผมลองให้ Claude ตัดวิดีโอให้หลายรอบแล้วแต่รู้สึกมันยังไม่ค่อยสวยถูกใจเท่าไร
Same pain, other posts
ลองทำหนังสั้นแบบความรู้ 0 รู้สึกปวดหัวมาก … มี SKILL ไหนบ้างสำหรับกำกับบทพูด และ Storyboard , มุมกล้อง
อยากทำสไลด์ให้ดูสวยงามขึ้นด้วย claude code พี่ๆมีเทคนิคหรือวิธีการที่ดีๆ เเนะนำไหมครับ
พี่ๆสายใช้ Claude สร้าง/ทำ Animation หารูปแบบ,ตัวอย่าง,แนวทาง,ไอเดียใหม่ๆ จากเว็บไหนกันบ้างครับ ใครมีช่วยแนะนำหน่อยค้าบ
ถ้าอยากให้ Claude วาดรูปอนิเมชั่นสวยๆ ต้องใช้สกิลอะไร หรือทำยังไงหรอครับ?
Three business opportunities
OPPORTUNITY 01
Claude video-editing skill pack with Thai 'pro' grammar
- What
- A packaged set of Claude skills driving ffmpeg/Remotion that enforces shot pacing, Thai subtitle typography, safe margins, loudness targets and export presets, so output looks pro without 10 failed attempts.
- Who pays
- Freelance editors who lose clients on a cut that 'isn't pretty', and creators shipping daily who pay per saved iteration.
- Price
- hypothesis: ฿990 one-time pack, ฿290/mo for updated presets
- Why now
- Claude Code can now drive a real render pipeline, but nobody has encoded what a polished Thai short-form cut looks like - the gap he hit.
- Effort
- M
- Moat
- Curated Thai subtitle/typography and pacing presets plus accumulated data on which edits get rejected.
- Risk
- CapCut already solves aesthetics for free and is the default answer for Thai creators.
OPPORTUNITY 02
Pre-flight 'ugliness score' QA on rendered video
- What
- Upload a rendered cut and get a score for dead air, jittery cuts, subtitles off-safe, low-res frames, audio level and beat mismatch, with an auto-fix list Claude can apply before export.
- Who pays
- Creators and studios shipping volume who physically cannot watch every frame before posting.
- Price
- hypothesis: ฿0.50 per rendered minute, or ฿490/mo flat for a channel
- Why now
- AI-assisted editing multiplied output volume; review, not generation, is now the bottleneck his post describes.
- Effort
- M
- Moat
- A labelled corpus of 'good vs rejected' Thai short-form edits that trains the scoring thresholds.
- Risk
- Aesthetic judgment is subjective; false positives destroy trust and it's hard to charge for a score.
OPPORTUNITY 03
One-time pipeline setup retainer ('make your Claude editor produce pro cuts')
- What
- A productized setup: an editor configures the client's asset folders, timeline templates, subtitle style and render presets around Claude plus a studio tool, then trains the team.
- Who pays
- Mid-size Thai channels and brand content teams with budget but no technical editor.
- Price
- hypothesis: ฿25,000 per setup engagement
- Why now
- He personally went through many failures to find the working stack; most teams will pay to skip that discovery.
- Effort
- S
- Moat
- Reusable setup playbooks that feed directly into the skill pack and QA tool above.
- Risk
- One-off revenue with no compounding; clients expect ongoing free support.