Cattle auction prices in Python

USDA AMS publishes cattle auction reports as PDFs and text. Our API turns them into one row per lot, so you can pull prices into a Python script or notebook in a few lines.

With requests

import requests

r = requests.get(
    "https://api.kossinlabs.com/v1/prices",
    params={"state": "NE", "class": "STEERS",
            "weight_min": 500, "weight_max": 700, "limit": 100},
    headers={"X-API-Key": "kl_free_demo_2026"},
    timeout=30,
)
r.raise_for_status()
lots = [l for l in r.json()["prices"] if "Cwt" in (l["unit"] or "")]
head = sum(l["head_count"] for l in lots)
avg = sum(l["head_count"] * l["price_avg"] for l in lots) / head
print(f"{len(lots)} lots, {head} head, avg ${avg:.2f}/cwt")

kl_free_demo_2026 is a free demo key for trying it out. Lots come back newest first.

With pandas

Add format=csv and read it straight into a DataFrame. Set a User-Agent: Cloudflare blocks Python's default urllib one, so a bare read_csv(url) gets a 403.

import pandas as pd

url = ("https://api.kossinlabs.com/v1/prices?state=NE&class=STEERS"
       "&weight_min=500&weight_max=700&limit=200"
       "&format=csv&api_key=kl_free_demo_2026")
df = pd.read_csv(url, storage_options={"User-Agent": "my-cattle-script"})
df = df[df["unit"].str.contains("Cwt", na=False)]

daily = (df.assign(value=df.head_count * df.price_avg)
           .groupby("report_date")[["head_count", "value"]].sum())
daily["avg_cwt"] = (daily.value / daily.head_count).round(2)
print(daily[["head_count", "avg_cwt"]].tail())

That gives a head-weighted average $/cwt per sale date. Weighting by head keeps a 5-head lot from counting as much as a 300-head one.

What one lot looks like

A real row from Bassett Livestock Auction (NE) on Nov 11, 2026:

{ "auction_name": "Bassett Livestock Auction", "state": "NE",
  "report_date": "2026-11-11", "category": "FEEDER CATTLE",
  "class_name": "STEERS", "grade": "Medium and Large 1",
  "unit": "Per Cwt / Actual Wt", "head_count": 366,
  "weight_min": 500, "weight_max": 549, "weight_avg": 528,
  "price_min": 445, "price_max": 495, "price_avg": 476.47 }

Rough lot value is (weight_avg / 100) * price_avg * head_count. Check the unit first, because some replacement lots are priced per head. More on that: cattle price per cwt.

Filters

Full list and the OpenAPI spec: api.kossinlabs.com. Prefer a spreadsheet? See cattle prices in Google Sheets.

Your own key

The demo key is shared. Get your own free key in one click, or Starter is $29/mo for higher limits: subscribe. 30-day money-back guarantee. Product page: cattle auction prices API. Source: USDA AMS Livestock, Poultry & Grain Market News (public). We are not affiliated with USDA.