snips-nlu

Why we included this project

Snips NLU is a self-contained Python library that turns free-form sentences into structured commands without needing a hosted API. It handles two steps in one pass: it figures out the user's intent, like recognizing that a sentence is a request to turn on a light, then pulls out the parameters that matter, such as the room or the time, and normalizes values like dates into clean ISO formats. Because the engine trains on your own labeled data, you define the intents and slots instead of being locked into a vendor's schema, and it runs locally with modest memory use. That makes it a good fit for on-device or privacy-sensitive assistants, and the Python API plus command-line training workflow keep prototyping quick. It is also a useful reference for seeing how intent classification and slot filling fit together in a real pipeline.