Released after
the meeting.
the meeting.
SAPIENSQ · MIRROR REAL ONES · PRODUCT 1
We don't forecast rates.
We simulate the committee that decides them.
Built for those looking to respond quickly to policy changes.
Each meeting is reconstructed from the public record, with named members, rotating seats, and prior debates.
The model returns the committee’s next decision.
FOMC is our first case.
The April 29 call
Locked before the meeting.
The FOMC statement releases on April 29, 2026 at 14:00 ET.
Until then, this card does not change.
The actual outcome appears beside it once the Fed publishes.
What's inside
Three parts. One room.
Each part appears in the deliberation transcript.
Request access to follow the process turn by turn.
01 · Persona
Who is in the room.
Each member is built from their statements, votes, and speeches. Seats refresh every meeting. Rotations and new appointments land where they should.
Where to look. Each output traces back to the speeches and votes that support it.
02 · Memory
No look-ahead, ever.
Every source document carries a release date. Each meeting reads only what was public at the time. Sources include Beige Books, minutes, statements, and transcripts.
Where to look. Citation tags link each turn to its source and release date.
03 · Deliberation
A structured deliberation, end to end.
The process runs in four phases. Briefing, debate, draft, and vote. The draft is grounded in the policy regime of that period, not a generic baseline.
Where to look. The draft’s reasoning chain is visible step by step.
Recent calls
What the Committee decided, and what we called.
Recent meetings where direction and magnitude matched.
N=3 reproducibility · Gemini 3 Flash
Run it three times. Read what stays.
Three independent reruns over the same 17 meetings.
Mean results are shown below.
Per-run breakdowns are available on request.
94.1%
Direction AR
Vote-direction accuracy (CUT / HOLD / HIKE)
4.07
Statement LLM judge
3-judge ensemble (Claude · Gemini · GPT), scored on a 1–5 scale
3.37
Minutes LLM judge
3-judge ensemble (Claude · Gemini · GPT), scored on a 1–5 scale
83.6%
Vote AR
Per-member vote alignment with the real FOMC
0.820
Statement embedding cos
Semantic similarity to the real FOMC statement
0.727
Minutes embedding cos
Semantic similarity to the real FOMC Minutes