…
…
01Overview
The latest statement, scored four ways — every figure traced to its source.
Section weights
How much each part of a release moves the composite tone score. Loaded from the database, recalibrated against past FOMC moves.
See the full method →02Breakdown
The Committee’s next move, as a full distribution of probabilities.
Signal Divergence
03Data
Pick any table from the pills to see its chart and rows. Click any row for the full record.
Core CPI · Unemployment · 2Y Yield
04Method
How a Fed release becomes a forecast — four stages, every figure traceable back to the underlying data.
From the publication of a Fed statement to the projection on this page — the four stages that translate the Committee's language into a rate-move forecast.
For the curious
The three agents, how each release is scored section by section, why the weights live in the database, and the statistical model behind the forecast.
05Feed
Live ingestion target. FOMC statements flow through the analyst agent and accumulate here.
06Run it yourself
This page is the read-only deployment. The whole pipeline — scrapers, analyst, strategist — runs on your own machine.
Clone the repository and python run.py setup writes the environment, builds the database and starts the agents. Retune the analyst's section weights, ingest documents by hand, or point the scraper at the synthetic FakeFed fixture site to watch a statement travel end to end.