How it works

Every AssetFrame edition runs through the same pipeline — research, compile, register, publish, score, append. The qualitative work is done by an AI analyst; the numbers, the predictions and the scoring are done by a deterministic engine, so the parts you’re asked to trust are the parts that can be audited.

1. 01

   Research

   An AI analyst studies the instrument and writes the thesis, the scenarios and the catalysts that could move it — the qualitative view, in plain English.

2. 02

   Compile

   A deterministic Python engine turns that view into numbers: the key price levels, conditional long/short setups with risk:reward, and a calibrated confidence score. Same inputs, same output — every figure is reproducible.

3. 03

   Register

   Before the session opens, the engine logs falsifiable predictions — exact levels and an exact window. Each one can be proven right or wrong; nothing is left vague.

4. 04

   Publish

   The free Snapshot opens for everyone and the Pro report unlocks with a subscription, served from a CDN so it stays fast at any traffic. Both render from one canonical payload behind a strict QA gate.

5. 05

   Score

   After the window closes, the engine grades each prediction against the actual market — Hit, Miss or Not triggered — with no human nudging the result.

6. 06

   Append & learn

   Results land in an append-only ledger that's never edited or re-tuned. The ledger is also an input: the engine learns which setups and regimes have worked — with no look-ahead, since a call is only ever scored after its window.

The confidence score, in plain English

Each Pro report carries a confidence score from 0 to 100. It isn’t a hand-waved number: the engine blends three things — the market structure the setup is built on, the ledger’s own track record for similar calls, and how well the catalysts are sourced. Because it’s graded against the actual market after every window, it’s calibrated — the goal is that calls rated, say, 70 actually come true about 70% of the time. It is a calibrated estimate of how a setup may resolve, not a guarantee, a probability of profit, or a signal to trade. Always read it next to the risk rating and the prediction window.

Free vs Pro

Stay in the loop

For developers & AI agents