docs/source/index.md

Kimodo Documentation

Kimodo

Scaling controllable human motion generation

Get Started Interactive Demo GitHub Tech Report

Overview

Kimodo is a kinematic motion diffusion model trained on a large-scale (700 hours) commercially-friendly optical motion capture dataset. The model generates high-quality 3D human and robot motions, and is controlled through text prompts and an extensive set of constraints such as full-body pose keyframes, end-effector positions/rotations, 2D paths, and 2D waypoints. See the project page for details.

Highlights

Controlled Generation

Text prompts combined with full-body, root, and end-effector constraints.

Human(oid) Support

Model variations for both digital humans and humanoid robots.

Interactive Demo

Timeline editing, real-time 3D visualization, and example presets.

Quick links

• Installation
• Quick Start
• Command Line Interface
• Interactive Demo
• Project Structure

:maxdepth: 3
:caption: Getting Started
:hidden:

getting_started/installation
getting_started/quick_start

:maxdepth: 2
:caption: User Guide
:hidden:

interactive_demo/index
user_guide/cli
user_guide/constraints
user_guide/output_formats
user_guide/motion_convert
user_guide/seed_dataset
user_guide/configuration

:maxdepth: 2
:caption: Key Concepts
:hidden:

key_concepts/model
key_concepts/limitations
key_concepts/motion_representation
key_concepts/constraints
key_concepts/skeleton

:maxdepth: 2
:caption: Benchmark
:hidden:

benchmark/introduction
benchmark/pipeline
benchmark/metrics
benchmark/results

:maxdepth: 2
:caption: Reference
:hidden:

project_structure
project_info
api_reference/index