Posecode

Posecode is an open source format for describing human movement as structured text. Instead of opaque motion capture files or timeline based animation tools, you write movement instructions in a readable language that produces deterministic 3D animation. The same source code always generates the same motion, which makes it versionable, diffable, and auditable in ways that binary formats are not.

The use case it targets is anyone who needs to create, edit, or validate human motion data without relying on proprietary software. That includes game developers prototyping character animations, fitness apps that need to demonstrate exercises, rehabilitation tools showing correct movement patterns, and AI workflows where motion needs to be generated or verified programmatically. Because Posecode is text based, large language models can generate movement descriptions that humans can then review and refine.

The language itself includes commands for common movement operations. Ground lock preserves an existing floor contact point. Reach moves a limb toward a target. Pin shifts the whole body around an anchor point. Grip solves hand positions onto objects like bars or rails. These primitives combine to describe complex sequences while remaining readable. If you know what the commands mean, you can look at a Posecode file and understand what the figure will do without running it.

The Posecode Playground is the primary interface for working with the format. It provides a live editor on one side and a 3D animated figure on the other. As you type, the figure updates in real time with constraint checking that flags physically implausible poses. You can play animations at different speeds, browse a library of pre built movements for reference, and export results in BVH format for Blender and traditional animation tools or glTF and GLB for Three.js and web applications.

The project is structured with different licenses for different components. The core specification and tools use Apache 2.0, which is permissive. Product facing components use AGPL 3.0, which requires sharing modifications. Commercial licensing is available for organizations that need to embed Posecode in closed source products without the AGPL obligations.

What makes Posecode distinctive is the combination of inspectability and determinism. Motion capture gives you realistic movement but the data is opaque. Keyframe animation gives you control but the timeline is not self documenting. Posecode produces motion that you can read, edit, diff, and generate programmatically while staying grounded in physical constraints. For workflows where auditability or AI generation matter, that trade off makes sense.

The tool is free to use through the web playground, and the format itself is open for anyone to implement. If you are building motion related features and want a text native approach that integrates with version control and AI pipelines, Posecode offers a foundation that proprietary formats cannot match.