Making Videos with Claude: Opus 5.5 Animation and Finished-Film Examples

A low-cost short-video pipeline that uses no video generation model and no on-camera talent: Opus 5.5 researches and writes the script, TTS voices it, GPT Image 2.5 generates the assets, and the result is edited and published automatically.

A Japanese developer made a small "Chill Game" with Claude Opus 5.5 and shared a video of it, aimed at people worn out by AI and dopamine feeds.

The author says the entire Robotron 1715 model, about 21k triangles, was made by Opus 5.5 in Blender without a single manual action, and plans to use the assets in later videos.

The author shows a website built with Claude Opus 5.5 and human direction, aimed at the kind of visual quality that gets investors to open their wallets.

A GitHub repo gathers 475 viral videos made by having Opus 5.5 write code, each with its prompt, covering motion graphics, explainer animations, 3D scenes and mini games.

The open-source Lemo-Opuscar project lets Opus 5.5 render frames with canvas/WebGL, compose music and TTS narration, producing a 1080p short film in 30-60 minutes.

The author says Claude Sonnet 5.5 made the whole video in one shot, with every frame generated by code, and calls it a sign that humans are done for.

A roundup of ten AI video workflow posts on X with over 20,000 bookmarks combined, covering Opus 5.5 code-made videos, edit-free production, open-source prompts and storyboard templates.

The author says the video made with Opus 5.5 and Seedance 2.5 is unbelievable and shares the finished clip.

The author made a demo with Opus 5.5 that seems plain but is actually very impressive, showing the result in a video.

The author used Opus 5.5 to make an explainer video that runs from a bronze-gear computer built 2,000 years ago to today's AI, in what they call an AI knowledge Slopcore style.

Ryan Sael used Claude Opus 5.5 to turn a PDF report on flooding in Maranhão, Brazil, into a 3D scene in the browser with a draggable water-level slider.