Kite Platform API: augment robot datasets and train policies from your code

The Kite Platform API runs the platform from your own code. Today that means three things. Augmentation: point Kite at a teleoperation dataset, describe a visual change in plain language, and get back new episodes as a standard LeRobot dataset. Training: pick your policies and a GPU tier, and get back trained checkpoints. Both run on cloud GPUs you never have to manage, and the output of the first is valid input to the second. Evaluation (beta): send a Kite SDK spec and get your policy scored in simulation on Kite’s machines. See Evaluations.

Get started Explore the API

Augment

Point Kite at a dataset, describe a visual change in plain language, and get back new episodes as standard Parquet.

Train

Fine-tune VLA and imitation-learning policies on cloud GPUs. Train several in one call and compare them.

Evaluate — beta

Run your policy on a Kite SDK spec in simulation and get a scorecard, before you touch real hardware.

Start with the quickstart to run your first augmentation, or read authentication to create an API key. Every endpoint is listed in the API reference.

Prefer a terminal or an AI assistant? The same platform is available through the Kite CLI and the Kite MCP server, which share this API’s authentication.