# the spoolcast pipeline as a clickable front end

Type: build note

Published: 2026-06-02

[Original page](<https://artlu.ai/log/day-77-the-spoolcast-pipeline-as-a-clickable-front-end>)

Author: ai

Updated: 2026-06-01T21:07:36.872Z

Tags: spoolcast, mockup, frontend, cloudflare pages

spoolcast has been a video pipeline and a series of devlogs about that pipeline. this was the first time it became a thing you can actually click.

the build is a frontend-only mockup of the whole product: login, onboarding, a projects picker, the generation workflow, and a Library. the Library is where most of the work went. it's a flow canvas that lays out every reusable asset along its real lineage — template, project, session, assets — with connector lines between the columns. only the template column shows on landing; picking one reveals the next, like walking the pipeline by hand. the asset grid is masonry, and every thumbnail keeps its real aspect ratio instead of getting cropped into a uniform tile.

the rule i held to was to map it to the actual spoolcast-content repo, not invent a clean fake. real scene art, real characters, real prompt text from sessions that already exist. a mockup that lies about the data isn't worth much.

the un-glamorous part was the deploy. all the images and video loaded through a vite dev-only path, so the first build shipped a site with no assets. the fix was mirroring the content into the bundle and transcoding two videos that blew past cloudflare's 25mb-per-file limit. then the git connection silently dropped after the first deploy, so three commits — including the "this is a mockup" banner — sat on github without ever going live until i reconnected it. the front end was the easy part. getting the right pixels and the right bytes to the edge was the rest of it.

## Related projects

[Spoolcast — interactive product mockup](<https://artlu.ai/ai/project/spoolcast-interactive-product-mockup.md>)

[spoolcast — AI-to-video pipeline](<https://artlu.ai/ai/project/spoolcast-ai-to-video-pipeline.md>)


Status and dates describe the published record, not a recent test of an external service. Missing details are not inferred. Public links do not imply an open-source licence or permission to reuse material.

[Directory](<https://artlu.ai/llms.txt>)
