Skip to main content
The Kite Platform API lets you run the platform programmatically — today, that means generating augmented robotics datasets on demand. Point Kite at a dataset, describe the change you want, choose how many episodes, and Kite generates them on its GPUs and delivers a ready-to-train dataset.

​
Base URL

All requests go to:
Every request is authenticated with an API key sent as a Bearer token. See Authentication.

​
How it works

You provide four things; Kite does the rest. There’s no infrastructure to manage — generation runs on Kite’s GPUs, episodes are produced in parallel, and the finished dataset is delivered in standard LeRobot Parquet format.
1

Dataset

A Hugging Face LeRobot dataset to augment, e.g. lerobot/pusht.
2

Prompt

A plain-language instruction — the visual change to apply to every episode.
3

Episodes

How many augmented episodes to generate (1–50 per run).
4

Destination

Kite-hosted (download the Parquet files) or pushed to your own Hugging Face repo.

​
Asynchronous by design

Augmentations run asynchronously. Creating one returns immediately with an id; you then poll the run as it progresses through queued → processing → succeeded. A completed run gives you a standard LeRobot dataset: Parquet tables for states and actions plus MP4 camera video. It’s the same format Kite training accepts, so you can train on it directly.

​
Conventions

Full request and response schemas are published as OpenAPI at https://api.kiteml.com/v1/openapi.json, rendered at /v1/docs.
  • JSON everywhere. Requests and responses are application/json unless noted.
  • Resource ids are prefixed — an augmentation is aug_…, an API key is key_… — so they’re easy to recognize in logs.
  • Timestamps are ISO 8601 in UTC (e.g. 2026-07-21T09:14:00Z).
  • Idempotency is supported on run creation via an Idempotency-Key header — see Augmentations.