Show HN: Posecode, a text language and renderer for human movement

Posecode, a text language and renderer for human movement, has been released on Show HN. The language uses commands like ground-lock, reach, pin, grip, and cue to animate movements such as deadlifts or stretches, with an LLM prompt available to generate valid .posecode documents.

- 1 Copy the prompt. Click Copy LLM prompt top right . It teaches a capable model how to write the Posecode language. - 2 Ask an LLM for a movement. Paste the prompt, then ask for what you want: “write a deadlift” , “a hamstring stretch” , “cat-cow” . For a representable request, it should reply with only the raw .posecode document. - 3 Paste it here. Drop the raw .posecode text into the editor on the left. Valid documents animate; parse errors, range clamps, and unresolved reach targets appear below the editor. Authoring reference The language specification /spec.html is the normative Posecode contract. The LLM guide /llm-guide.html is a self-contained authoring aid built from that contract. ground-lock preserves an existing floor support; reach moves a limb; pin moves the whole body around one anchor; and grip solves hands onto a bar or rails. A cue is display-only coaching text and never changes the animation. The prompt you’ll paste into the LLM