adapt

Why we included this project

Adapt turns free-form user phrasing into a structured action your app can act on. It's a lightweight Python library that parses natural language and returns the detected intent, a confidence score, and tagged entities, so a phrase like "play my Joan Jett Pandora station" becomes a play-music command without any model training. Because matching is rule- and vocabulary-based rather than statistical, it stays fast on modest hardware and runs fully offline, which suits embedded devices, remote controls, and other constrained setups where a cloud round-trip isn't practical. The IntentBuilder API is easy to wire into your own assistant or skill layer, and the deterministic, explainable results are a good fit when you only need to handle a limited set of commands.