A working pipeline book: batch-generate with Wan 2.2 and HunyuanVideo 1.5 on a single RTX 4090, automate queue → captions → packaging, and decide "own GPU vs API" with benchmark data and license facts — not forum folklore.
Read chapter 1 free1 · Own Your Minutes — cost math (0.3 THB vs 14 THB per clip), where open weights honestly rank, the license map
2 · The 24GB Sweet Spot — own vs rent vs API, stable environments, your benchmark suite
3 · The Generation Core — seeds, resolution/frame/step trade-offs, batch + cost ledger
4 · The Production Pipeline — file-based job queue, retries, captioning, ffmpeg packaging
5 · Turning Clips into Products — what sells, platform AI-disclosure rules, pricing from API anchors, getting paid from Thailand
Every chapter ships with runnable companion scripts (benchmark suite, batch runner, caption packager, price calculator) — standard library plus diffusers, each with sidecar metadata on every artifact.
Get "Own Your Minutes" (the full cost + license chapter) as a PDF, plus the benchmark CSV template. No spam — build-log updates only while the book is being written.
Another AI video book?
No — it's a pipeline book. It doesn't teach prompt tricks or worship one model (they age in months). It builds the boring layer under any model: job queue, cost ledger, packaging, licensing per market.
Do I need a 4090?
It's the reference card. Anything ≥16GB VRAM runs the 480p configs; the book shows when renting (RunPod 4090 from ~$0.34/hr) beats owning, with a script to measure your own numbers.
Will it be out of date in six months?
Model names will change; the book is built for that — everything hangs off a job schema and swappable backends, and the cost/licensing *methods* stay valid. Buyers get updated companion code.
Code language?
Python. Standard library plus diffusers for generation, ffmpeg for packaging. Every script is standalone with --help.
Where do I buy it?
It's published as a work-in-progress book: buy now, get each chapter as it lands, price goes up at completion.
Chapter 1 free · companion code included · work-in-progress pricing
Start with chapter 1