SmarterX presents the first-ever AI model built for CPG compliance teams. Our API allows you to enrich large product catalogs with faster, smarter classifications, at scale.
When working with Smarter-1, you'll be dealing in Decisions, which are collections of classifications. For instance, the RCRA
decision returns a set of classifications that determine whether your products meet the RCRA requirements for controlled hazardous waste.
The complexity involved in calculating Decisions requires an asynchronous process, so you'll put in jobs and periodically poll for their status. We recommend waiting at least 30 seconds per product in your request, though some classifications will complete much faster. Expect variability: decisions that rely on an LLM will vary in processing time based on several factors.
You receive your completed Decisions in the status call itself.
Each job may request one or many Decisions against one or many products. To predict your credit consumption, the formula is credits_used = decision_count * product_count
. The maximum number of products per job is 5.
The Smarter-1 AI returns accurate classifications in many cases with absurdly little input, for instance, just product name and brand name. However, some classifications will benefit from more specific input data, such as flashpoint
.
The API's response payload provides detailed guidance on what caused the model to respond a certain way, so if you get an unexpected classification, experiment to see whether your inputs are insufficient. Or, just contact us. We'd be excited to explore the model together with you.
If you run into any issues, please contact us at our support site. Enjoy!