--- title: "Most AI video tools output slop. I built a deterministic Three.js + LLM pipeline instead" url: https://daily.dev/posts/most-ai-video-tools-output-slop-i-built-a-deterministic-three-js-llm-pipeline-instead-xq4ivoygt source_url: https://daily.dev/posts/most-ai-video-tools-output-slop-i-built-a-deterministic-three-js-llm-pipeline-instead-xq4ivoygt type: freeform source: "Varun Dev" author: "Varun Dev" published: 2026-06-24T10:11:41.457Z updated: 2026-06-24T10:11:59.796Z tags: ["webdev", "llm", "threejs"] reading_time: 2 upvotes: 0 comments: 0 language: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Most AI video tools output slop. I built a deterministic Three.js + LLM pipeline instead **[Varun Dev](https://daily.dev/sources/ebbuxpsgx5yyxfu0ghtny)** ยท [@var_dev_sh](https://daily.dev/var_dev_sh) ยท 2 min read ยท 0 upvotes ยท 0 comments ## Summary A developer built Unstill, a deterministic 3D video generation pipeline to avoid the frame-drift issues common in AI video generators. The architecture separates concerns: an LLM authors a pure-data scene file (no code generation), a unified Three.js + GSAP + Matter.js engine renders it for both live preview and video export, and Remotion handles MP4 export with TTS narration synced per chapter. The key technical insight is making all motion a pure function of timeline.time() rather than accumulated delta-time, enabling Remotion to parallelize rendering across workers with zero drift and ensuring pixel-identical output between preview and export. Quality is improved through structured-output schemas, domain accuracy packs, and a visual-critique self-correction loop. ## Content ๐Ÿ”ง Most "AI video generators" output slop with drifting frames. I'm a dev โ€” I wanted deterministic, frame-perfect 3D. So I built my own engine. Here's the architecture (with a teardown of an actual engine it generated ๐Ÿ‘‡). TL;DR: I built Unstill โ€” you give it a prompt, an LLM authors a pure-data scene, a custom Three.js + GSAP engine renders it interactively, and Remotion exports a narrated MP4. The hard part wasn't the LLM โ€” it was making the live preview and the exported video pixel-identical. How it works: 1. The LLM doesn't render โ€” it authors data. It outputs a file: pure data describing objects, tweens, camera moves, and chapters. No code-gen, no pixels โ€” so output is validatable, repairable, and version-controllable. 2. One engine, two outputs. A Three.js + GSAP + Matter.js engine renders that data for BOTH the live studio preview and the video export. The catch: "looks right in preview, broken in the MP4." The fix was making all motion a pure function of timeline.time() (no accumulated dt), so any frame renders identically and Remotion can parallelize the render across workers with zero drift. 3. Remotion for the MP4 โ€” with AI narration (TTS) synced per chapter, captions, and a coded like/subscribe/bell outro shared between preview and export so they never diverge. 4. Quality is a loop, not a prompt. Structured-output schemas + domain accuracy packs + a visual-critique pass that re-renders and self-corrects. Here's a 4-stroke engine it generated end-to-end ๐Ÿ‘‡ โ–ถ๏ธ [https://www.youtube.com/watch?v=TWUhj6Hmv3M](https://www.youtube.com/watch?v=TWUhj6Hmv3M) What's your thought on this? #threejs #webgl #javascript #typescript #react #nextjs #ai #machinelearning #webdev #devtools --- Tags: [#webdev](https://daily.dev/tags/webdev), [#llm](https://daily.dev/tags/llm), [#threejs](https://daily.dev/tags/threejs) [View this post on daily.dev](https://daily.dev/posts/most-ai-video-tools-output-slop-i-built-a-deterministic-three-js-llm-pipeline-instead-xq4ivoygt)