Pipelines
Everything on this site came through some version of what's below. It isn't secret — the value isn't in knowing the steps, it's in having built the bridges between them so they run without babysitting every frame.
Four phases, borrowed from film
01
Development
Story/Audience/Deliverables/Formats
What we are making, who it is for, what formats the deliverables need. All of it decided before a single token is spent — the one-page brief everything downstream points back to.
02
Pre-production
Model selection/Storyboards/Reference sheets/LoRA training/Music
Which models serve the project, and what they cost, shared openly. Storyboards iterated fast and locked. Custom models trained and approved. Music often gets made here too. Where token management pays off most.
03
Production
Generation/Dailies/Approve / re-shoot
Footage generated in ComfyUI from the locked references, master soundtracks built. Reviewed daily, approved or re-shot. Time and budget gate every generation, so the production never goes off the rails.
04
Post-production
ComfyUI → Resolve bridge/Edit/VFX / titles / sound/Grade/Render
Edit, VFX, titles, sound, atmosphere, colour grade, render. Rough cut, pre-final, final.
When the tool doesn't exist, or exists and fights me, I write my own. Two that earn their keep:
For DaVinci Resolve
Cuts to the beat, without counting on your fingers.
Cutting video to music is harder than it should be. Master Beater takes an MP3 or WAV, tracks it into beats and measures, reads your project frame rate, and renders an MP4 you drop on your timeline as a visual beat cue. It also exports a subtitle track that calls out the beat and measure as you scrub. Song is 118 BPM and you need 120? A converter panel retimes the track and re-exports the cue at the new tempo.
In: MP3 / WAV · Out: MP4 beat cue + subtitle track · Reads project frame rate · Tempo converter · Used on: [DANO: which projects]
View on GitHub →In development
For ComfyUI, DaVinci Resolve
Fewer round trips between ComfyUI and Resolve.
What comes out of one is rarely what the other wants. Dano Nodes are the custom nodes I am building to make the round trip between the two easier, or remove it. Currently four print-look image nodes — Risograph, Risograph Simple, Halftone Simple, Halftone Complex — pure numpy, CPU-friendly. [DANO: confirm these are the ones to feature, and whether more have shipped since.]
View on GitHub →Interactive
Not every deliverable is a video. The same discipline applies to interactive e-learning built in Articulate Storyline, steering AI to produce every slide. See it on thee-learning project page.
Tokens aren't free
[DANO: honest principles or numbers — where the money goes in an AI production and how the pipeline keeps it in check.]