Developer docs

Two free JSON APIs. No key, no signup — every read endpoint is open. Data is 100% USDA; when a source can't deliver, the API says so instead of guessing.

Quickstart

Pick a base URL and fetch. Everything is GET and returns JSON.

curl "https://ag.morningbrief.link/v1/grain/prices?product=corn&geography=IL"

API	Base URL	What it covers
Magpie	`https://ag.morningbrief.link`	USDA market data: prices, forecasts, export sales, trade flows, crop conditions
Farmland	`https://land.morningbrief.link`	USDA land values and cash rents, all 50 states

Magpie — USDA market data

Endpoint	What you get
`GET /v1/grain/prices`	Daily grain bids (corn, soybeans, wheat, …) by region
`GET /v1/grain/history`	Price time series for a product and region
`GET /v1/livestock/prices`	Auction and direct-trade livestock prices
`GET /v1/dairy/prices`	Weekly dairy product prices
`GET /v1/fertilizer/prices`	Fertilizer prices (urea, ammonia, …)
`GET /v1/nass/condition`	Weekly crop condition ratings + good/excellent share
`GET /v1/nass/progress`	Planting and harvest progress by stage
`GET /v1/nass/stats`	Yields, production, acreage
`GET /v1/wasde/projections`	Monthly USDA supply & demand forecasts
`GET /v1/wasde/history`	How past forecasts were revised
`GET /v1/esr/sales`	Weekly export sales contracts by country
`GET /v1/esr/history`	Weekly sales history
`GET /v1/gats/flows`	Monthly export shipments by destination
`GET /v1/gats/compare`	Contracts vs. actual shipments gap
`GET /v1/weather/current`	Current conditions for growing regions
`GET /v1/weather/forecast`	Forecast for growing regions
`GET /v1/status`	Dataset row counts and last fetch

Every response carries a _provenance stamp: where the data came from, how often it refreshes, and the exact date it covers. Gaps are labeled, not hidden: no_data (nothing matches), source_unavailable (upstream failed), withheld (excluded by a quality check, with the reason).

Farmland — land values & cash rents

Endpoint	What you get
`GET /v1/land/values`	Farmland values by state/county and year (?state=IL&year=2026)
`GET /v1/land/rents`	Cash rents by state/county and year
`GET /v1/status`	Dataset coverage and row counts

Source: USDA NASS. Updated once a year when NASS publishes (August).

MCP — for AI agents

Magpie speaks the Model Context Protocol. One endpoint, standard JSON-RPC:

POST https://ag.morningbrief.link/mcp
Content-Type: application/json

{"jsonrpc":"2.0","id":1,"method":"tools/list"}

Protocol version 2024-11-05. Every tool result arrives in an envelope — {status, data, provenance} — so agents can tell fresh data from a gap without guessing.

26 tools

Tool	What it does
`get_grain_price` / `get_grain_history`	Grain prices, latest or time series
`get_spreads`	Cash price spreads
`get_livestock_price` / `get_livestock_history`	Livestock prices, latest or time series
`get_dairy_price`	Dairy product prices
`get_fertilizer_price` / `get_fertilizer_history`	Fertilizer prices, latest or time series
`get_nass_stat` / `get_crop_stats`	NASS statistics (yield, production, acreage)
`get_crop_condition` / `get_crop_progress`	Crop ratings and planting/harvest progress
`get_wasde_projection` / `get_wasde_history`	USDA supply & demand forecasts
`get_export_sales` / `get_export_sales_history`	Weekly export sales by country
`get_trade_flows`	Monthly export shipments
`compare_contracts_vs_shipments`	Contracted vs. shipped gap
`get_weather` / `get_forecast`	Growing-region weather
`list_grain_products`, `list_livestock_products`, `list_dairy_products`, `list_fertilizer_products`	What each price dataset tracks
`list_wasde_commodities`, `list_esr_commodities`, `list_gats_commodities`	What each trade/forecast dataset tracks

How we handle missing data

A gap is a data problem, not an excuse to guess. When a dataset has nothing to show, the API tells you which kind of gap it is:

Signal	Meaning
`no_data`	Valid question, nothing matches the filter.
`source_unavailable`	The USDA source (or our copy) failed right now.
`withheld`	Deliberately excluded by a quality check — e.g. butter (USDA publishes only price differences, no base price) or rice trade flows (source unit errors). The reason ships in the response.