Non-autoregressive System 1 decision engine. Typed choice, score and yes/no decisions over any text in a single forward pass, in 100+ languages, with a router that picks the right checkpoint per request.
Python calibrationclassificationdecision-modelhuggingfacejevmodernbertmultilingualnlppythonpytorchroutingtyped-decisionszero-shot
docs(mcp): list laya_predict_long among the tools remote mode refuses
Rishit Sharmaf943f46e6b33
diff --git a/docs/cli-mcp.md b/docs/cli-mcp.mdindex b9d8183e..d28ecdd7 100644--- a/docs/cli-mcp.md+++ b/docs/cli-mcp.md@@ -353,8 +353,10 @@ Each Router keeps the value resolved at construction; environment changes requiror a restarted MCP process. An explicit request header also takes precedence over a UA set ona custom urllib opener; use this setting instead of that workaround.-`laya_shortlist`, `laya_tournament`, and prediction hook overrides return `unsupported_remote`, since these-paths run in the model process. For shared-server tournament selection, enable+`laya_shortlist`, `laya_tournament`, `laya_predict_long`, and prediction hook overrides return+`unsupported_remote`, since these paths run in the model process. The HTTP server has no long-state+scan, so for a state past one window run the MCP server without `LAYA_BASE_URL`; `laya_predict`+over HTTP reads the first window only. For shared-server tournament selection, enable[`LAYA_TOURNAMENT_GROUP` on the HTTP server](http-api.md#large-choices-laya_tournament_group) and call`laya_predict`: HTTP keeps probabilities for every original label (zero for eliminated labels),reports `x_tournament`, and counts usage across rounds. The in-process `laya_tournament` returns