Resources
Applied ML writing, the full documentation, and what other teams have actually built.
Notes on classification, benchmarks we've run, and what we've learned operating models in production.
Some of it is practical — how much training data you actually need, how to think about confidence thresholds. Some of it is research we ran on our own production data and published, including how well LLM confidence scores hold up under scrutiny.
Guides, core concepts, and the complete API reference.
Two ways in. If you'd rather learn by clicking, the Get Started path walks you through building a function in the console with no code. If you'd rather see the API, the quickstart gets you a live prediction in five calls.
How Nyckel gets used sector by sector.
Marketplaces, retail, digital asset management, ad tech, and others. Useful if you'd rather start from a business that looks like yours than from a capability.
What other teams built, and what it took.
Product tagging for an eCommerce catalog, moderation across images and text, spam identification, intent detection on sales replies, classification running on IoT devices.
Rapidly develop and deploy custom ML models that are accurate, secure, and easy to integrate.