Available on PyPI

Pose CLI — open motion analysis for serious work.

A pip-installable toolkit that turns 3D skeleton data into actionable sports analytics. Built for sports scientists, coaches, and performance analysts — with 20+ built-in tools, sport-specific plugins, and a local backend that runs entirely on your machine.

Get started Research collaboration

$ pip install pose-platform
$ pose analyze data/session.csv --tool stats
→ data_quality: good, tracking_coverage: 0.98

$ pose serve
→ http://0.0.0.0:8000
→ LAN  →  http://192.168.1.x:8000

$ pose annotate serve data/session.csv
→ Web annotator at http://localhost:8001

Built for the people who study motion

Three groups we hear from most.

Sports scientists & researchers

Run reproducible analysis across sessions and athletes. Export markdown / JSON for papers and reports. Publish your tools as independent pip packages.

Coaches & performance analysts

Evaluate technique, track progress, compare athletes. Sport-specific tools for badminton, sprint, and more — install only what you need.

PE programs & labs

Standardized motion statistics for student cohorts. Batch processing, CI-friendly JSON output. Zero cloud dependency — all analysis runs locally.

Key features

Everything you need to go from raw pose data to actionable insights.

🏃 Sports as Plugins

Built-in freeplay and sprint. Install sport-specific packages: pip install pose-platform[badminton]. Create your own with pose sports create.

🖥️ Built-in Local Backend

pose serve turns your computer into a full backend for Pose Coach — auto-upload, analysis, AI chat, training plans. No cloud required.

🏷️ Web Annotation Tool

Label recordings in your browser with synchronized video + 2D skeleton overlay. Create ground truth for calibration and algorithm validation.

🤖 LLM-Powered Analysis

pose chat lets you ask natural-language questions about your data. Defaults to DeepSeek (free tier). Also supports Claude and Qwen.

📊 20+ Analysis Tools

Stats, motion segmentation, action detection, velocity profiling, gait analysis, highlight extraction — all with a consistent CLI interface.

🔒 100% Local-First

Your data never leaves your computer. Analysis, AI chat, and the local backend all run on your machine. Cache built-in for instant re-runs.

Architecture

A local-first Python CLI with pluggable sports and a built-in backend.

20+ pluggable analysis tools — stats, motion segmentation, action detection, velocity profiling, and more
Sports as plugins — built-in freeplay + sprint; install badminton, basketball, or create your own via Python entry_points
Built-in lightweight backend — pose serve turns your computer into a full-featured backend for Pose Coach; analysis, AI coach chat, training plans, and reports all run locally
Web annotation tool — pose annotate serve opens a browser UI with synced video + skeleton for labeling
LLM chat — pose chat with DeepSeek (default, free tier), Claude, or Qwen
JointData3D format — single skeleton format shared with Pose Coach; accepts CSV and .bin
Runs 100% locally — no server required for analysis commands

Data format

Bring your own pose data — Pose CLI works with our compact 12-column CSV.

timestamp, joint, wx, wy, wz,
mx, my, mz,
ldq_x, ldq_y, ldq_z, ldq_w

Binary .bin from Pose Coach is also accepted natively. Output as Markdown reports or CI-friendly JSON. Works with ARKit, OpenPose, MediaPipe, and motion capture data.

Ready to put your motion data to work?

Pose CLI is available on PyPI — install and start analyzing in minutes. Drop us a line and tell us what you want to study.

Get started Research collaboration