Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.kynasmith.dev/llms.txt

Use this file to discover all available pages before exploring further.

​
Get started

Kynasmith is a movement-detection API. You author movement definitions in MoveSpec YAML, then run detection sessions that analyze pose data in real time and emit typed movement events. There are two official SDK paths:
  • Python SDK (kynasmith) for server-side, replay, and local camera workflows
  • Web SDK (@kynasmith/web-sdk) for browser-side live and replay workflows
You can also use the direct HTTP API from any language.

​
Evaluate first

If you want to understand the product before setting up credentials, start here:
  • Run the guided guest demo in the Workbench
  • Browse released public MoveSpecs in the catalog
  • Inspect the runtime event model in the API reference
The guest Workbench is the fastest way to see validation, replay execution, and typed outputs before you commit to the full integration path.

​
When you are ready to integrate

1

Log in to the portal

Go to app.kynasmith.dev and sign in or create an account.
2

Create an organization

After signing in, create your organization. This is the billing and team boundary for all your projects.
3

Create a project

Inside your organization, create a project. Projects scope your MoveSpecs, detection sessions, and credentials.
4

Create an API key

Navigate to Settings > API Keys > Create Key. Copy the key ID and secret — you will need them to authenticate SDK and API calls.
5

(Browser integrations) Set up a token endpoint

If you plan to use the Web SDK, build a backend endpoint that exchanges your API key for a short-lived access token. See Authentication for details.

​
Prerequisites

  • A Kynasmith evaluation path completed, or a clear target MoveSpec you want to run
  • A Kynasmith project created in the portal
  • A released MoveSpec version to run at detection time
  • Project-bound API key credentials
  • A backend token endpoint if you will call Kynasmith from the browser

​
Choose an integration path

Python SDK — Use when your application runs on a server, worker, desktop process, or scripted local tool. See the Python SDK quickstart. Web SDK — Use when your application runs in a browser and you want the SDK to own the MediaPipe browser runtime. See the Web SDK quickstart. Direct HTTP API — Use when you need to integrate from another language or prefer direct API calls. See the API reference.

​
Integration flow

1

Create and release a MoveSpec

Author MoveSpec YAML that defines the movements you want to detect. Validate the draft, create an immutable version, and release it. See MoveSpec.
2

Create a detection session

Start a live or replay session against the released MoveSpec version.
3

Stream pose frames

The SDKs keep raw video local. They extract pose landmarks with MediaPipe and send only pose frames to Kynasmith.
4

Read typed events

Consume event_batch, session_update, and session_done messages until the session reaches a terminal state.