Price terms as published
Each line is the vendor’s own wording, split from the price capture checked on 2026-10-27. Nothing here is rounded, converted or re-worded — the full sentences stay in the review below. Lines that carry only a price continue the plan named on the line above, in the vendor’s own order.
| # | Plan / term | Checked |
|---|---|---|
| 1 | Free — MIT licence for the code | 2026-10-27 |
| 2 | the song itself belongs to its original artists (credited in the repo) | 2026-10-27 |
One-liner: pdoom-video is a complete music-video project for “I’m Upping My P(doom)” — every frame is a deterministic function of song time, so the live browser preview and the offline export are identical, and the whole production pipeline (concept, treatment, audio analysis, renderer, scenes) was worked out with Claude Opus 5.5 in Claude Code.
This entry is compiled from the project’s GitHub repository and README, read on 2026-10-27. It is not a hands-on test: the video was not re-rendered for this page, no score is assigned, and every number below is sourced inline.
Verified facts (GitHub API and README, captured 2026-10-27)
| Attribute | Value |
|---|---|
| Repository | mexicat/pdoom-video |
| Stars | 3,539 (GitHub API, 2026-10-27) |
| Created / last push | 2026-09-24 / 2026-10-05 |
| Licence | MIT (GitHub API licence field) |
| Price | $0 — no paid tier, no hosted service |
| Output quality | 1080p60, or 4K60 for the offline export (per README) |
| Homepage | None (repo only; the video is on YouTube, linked from the README) |
| Song ownership | Not theirs — credits in the repo |
The star count comes from the GitHub API on 2026-10-27. The render-quality and production-process claims are quoted from the README.
What it is
This is one finished artifact rather than a toolbox: a generative, code-rendered music video with word-synced karaoke typography. The README explains the deterministic design — each frame is a function of song time, so the live preview in the browser and the offline render are pixel-identical. That property is what makes the project reproducible instead of a lucky one-off.
The production pipeline is documented in two files: docs/TREATMENT.md (the concept and style bible) and docs/ENGINE.md (the engine and scene API). The README also describes the audio-analysis work that produced the timing data: Demucs stem separation, CTC forced alignment cross-checked with Whisper, and beat/downbeat/onset analysis — all of it committed into data/lyrics.json and data/audio.json.
How it was made
Per the README, the video was made with Claude (Opus 5.5) in Claude Code: the concept and treatment, the lyric alignment and audio analysis, the renderer, every scene and the renders themselves were all worked out in conversation with Claude. The song is a “Claude-Pop” version credited in the repo.
Why deterministic rendering matters here
Most AI-video tools produce different frames on every run; this project deliberately does the opposite, and that choice is what makes it a teaching artifact. Because each frame is a deterministic function of song time, the compositor can play the video live in the browser at full quality and export the same footage offline — there is no “best render” to chase, and no prompt-drift between preview and delivery. It also means the project is fully auditable: run it once and you have the definitive version, which is why the README can confidently point to a specific offline quality target (1080p60 or 4K60) instead of promising a model’s average output.
Install & run
- Clone
mexicat/pdoom-video. - Follow the “Render the video” section in the README (the project ships its committed analysis data, so the render does not depend on re-running the audio pipeline).
- Watch the live preview in the browser, or export at 1080p60 / 4K60 offline.
Price & license
Free. The code is MIT-licensed; the song is not theirs and is credited to its original artists in the repo. There is no paid tier and no hosted service.
Alternatives
- Lemo-Opuscar — 59 ready-to-use styles with prompts and demo films; the reusable toolbox version of the same approach.
- mg-styles-15 — 15 motion-design styles, each with full prompt and source; more polished examples, less deep.
- awesome-opus5-5-videos — 703 community prompts behind viral Opus 5.5 videos (its README links to the commercial Skillry site).
- Video-model generators (Sora, Veo, Runway, Kling) — photorealistic footage via diffusion models, paid, non-deterministic.