hindcast

MCP servers

MCP server for commodity data (EIA · USDA · Brazil agro)

Commodity research runs on a calendar. An EIA inventory number on Wednesday morning, a WASDE at a fixed hour on a fixed day, a harvest estimate that arrives when the agency says it does — and prices respond to the difference between the release and what was expected, not to the eventual revised truth.

This server exposes those releases as tools where the release itself is the unit. An agent can ask for the latest published figures, for a specific past vintage, or for the difference between two consecutive releases, and every response carries the publication timestamp that makes event studies possible.

Brazilian agricultural data sits alongside the US agencies rather than in a separate silo, which is the correct arrangement for markets where a South American harvest and a North American forecast are two halves of the same question.

hindcast — session

preview

hindcast> compare the last two monthly estimates for a crop

→ tool: wasde_release { commodity:"...", release:"latest", diff:"previous" }

changed line items with both vintages side by side

source: USDA WASDE releases, each retained as published

Preview of the shape of an answer — endpoints open to the waitlist first.

What's covered

Join the waitlist — this page's dataset is already selected.

Need this one?votes for Commodities & Agro
Optional — helps us build in the right order. All skippable.

Role

Company

Datasets you need

Questions

Can I reconstruct what was known before a release?

Yes — that is the main use. Query as of the day before and the tool returns the prior vintage, which is what the market was positioned on when the new report landed.

Does it include futures prices?

Prices come from the market data catalog rather than this server, but they resolve to the same contract identifiers, so pairing a release with the price reaction is a join and not a project.

Join the waitlist. Vote the roadmap. First in, first served.

← Back to the data catalog