The Input Validator API gives you access to standardized product, variety, and tank mix databases. It also matches the raw product names found in machine files against known products, so you can identify what was actually applied or planted in a field operation.
How it works
Leaf maintains databases of agricultural products sourced from Agrian, CDMS, and John Deere. When a machine file contains a product name like “ams” or “counter,” the matching endpoint finds the closest known product and returns it with a confidence score.
Matching has two statuses:
PREDICTED — Leaf’s best guess, returned with a numeric score. Higher scores indicate stronger matches.VALIDATED — A human has confirmed the match is correct, either by approving the prediction or by manually assigning a different product.
You can approve a prediction or override it with a different product ID using the PATCH endpoint. Leaf tracks the full change history for each match.
Key concepts
Products — Chemical, fertilizer, or additive records with registration numbers, active ingredients, labels, and physical state. Products come from three label providers: Agrian, CDMS, and John Deere. Agrian and CDMS products are available globally; John Deere products are scoped to the Leaf user level.
Summarized products — Product names extracted directly from a Leaf user’s machine files, before any matching. Useful for seeing exactly what names the monitor recorded.
Varieties — Seed variety records with crop type, company name, and status. Currently sourced from John Deere at the Leaf user level.
Tank mixes — Predefined combinations of products with a carrier and one or more components, including solution rate and target crops. Sourced from John Deere.
Product matching — Given a field operation ID, Leaf returns the best product match for each raw product name in that operation. You can validate matches, override them, and view the change history.
What to do next
Last modified on March 24, 2026
