Skip to content

Commit c493616

Browse files
Merge branch 'main' into codex/issue-60-single-foot-ground-lock
2 parents 3fd801b + 2d31824 commit c493616

12 files changed

Lines changed: 3334 additions & 127 deletions

README.md

Lines changed: 40 additions & 17 deletions
Original file line numberDiff line numberDiff line change
@@ -62,6 +62,46 @@ Ask an LLM to explain physical movement and it usually returns unstructured pros
6262
</sub>
6363
</p>
6464

65+
For example:
66+
67+
> Bend your knees, move your hips backward, and keep your chest upright.
68+
69+
A human may understand that instruction, but a renderer cannot reliably determine:
70+
71+
- which joints should move,
72+
- by how many degrees,
73+
- in which coordinate frame,
74+
- over what duration,
75+
- in what sequence,
76+
- or within which physical limits.
77+
78+
Large language models can often reason about the components of human movement, but they lack a standardized syntax for expressing that reasoning in a renderable and testable form.
79+
80+
Posecode provides that missing representation.
81+
82+
### See Posecode in 28 seconds
83+
84+
From an LLM prompt to editable Posecode, validated 3D rendering, MCP tools,
85+
and a one-script web embed.
86+
87+
<p align="center">
88+
<a href="docs/launch-media/posecode-cut2-builder-16x9.mp4">
89+
<img
90+
src="docs/launch-media/posecode-cut2-payoff.png"
91+
width="800"
92+
alt="Watch the 28-second Posecode builder demo"
93+
/>
94+
</a>
95+
</p>
96+
97+
<p align="center">
98+
<a href="docs/launch-media/posecode-cut2-builder-16x9.mp4"><b>▶ Watch the 28-second builder demo</b></a>
99+
<br />
100+
<sub>Plain text in. Smooth, programmable 3D motion out.</sub>
101+
</p>
102+
103+
### Movement examples
104+
65105
<table align="center">
66106
<tr>
67107
<td align="center">
@@ -82,23 +122,6 @@ Ask an LLM to explain physical movement and it usually returns unstructured pros
82122
</tr>
83123
</table>
84124

85-
For example:
86-
87-
> Bend your knees, move your hips backward, and keep your chest upright.
88-
89-
A human may understand that instruction, but a renderer cannot reliably determine:
90-
91-
- which joints should move,
92-
- by how many degrees,
93-
- in which coordinate frame,
94-
- over what duration,
95-
- in what sequence,
96-
- or within which physical limits.
97-
98-
Large language models can often reason about the components of human movement, but they lack a standardized syntax for expressing that reasoning in a renderable and testable form.
99-
100-
Posecode provides that missing representation.
101-
102125
---
103126

104127
## Why Not Diffusion Text-to-Motion?
5.46 MB
Binary file not shown.
5.1 MB
Binary file not shown.
442 KB
Loading

0 commit comments

Comments
 (0)