Construction

constraints in, portfolio out

Nothing on this screen is advice. You state the constraints, the engine builds a portfolio from them, and the constraints stay next to the result so you can check one against the other. The language model writes the constraints; it never picks a name and never computes a weight.

Market scope

IDX

Construction cannot run for US yet

This is a data gap, not a decision about US. The form is hidden rather than left working: its AI step would still charge credits to write constraints nothing can be built from, and the build button would come back empty.

The share sleeve ranks on nightly factor scores, and a score is a percentile inside one market — a PER measured against Jakarta says nothing about US. We hold those scores for IDX and for no other market yet, so there is nothing to rank.
The bond sleeve is Indonesian government paper (SBN). There is no US equivalent wired up, and a government yield is not interchangeable across countries.
Sizing assumes the Jakarta lot of 100 shares and a rupiah budget throughout. Applied to a market that trades single shares in another currency, it would return positions that are wrong by a lot size and a currency at once.

Construction

constraints in, portfolio out

Market scope

IDX

Construction cannot run for US yet

The share sleeve ranks on nightly factor scores, and a score is a percentile inside one market — a PER measured against Jakarta says nothing about US. We hold those scores for IDX and for no other market yet, so there is nothing to rank.
The bond sleeve is Indonesian government paper (SBN). There is no US equivalent wired up, and a government yield is not interchangeable across countries.
Sizing assumes the Jakarta lot of 100 shares and a rupiah budget throughout. Applied to a market that trades single shares in another currency, it would return positions that are wrong by a lot size and a currency at once.