MicroLLM — Decision Studio (System One Engine)

Input Payload (/v1/systemone)

State (Text or JSON) No-Code Visual Question Builder (1-Click Append)

Typed Questions (noul, choice, score, multi) StartLux Cascade Escalation Threshold (g ∈ [0, 1]) Optional Reference Context (For Guardrail Groundedness / Hallucination Check)

ONNX INT8 Profile Registry (microllm pull) ~/.cache/microllm

Click any profile above or run microllm pull fast-cpu-int8 in your terminal.

Export to Code (Ready-to-Run Snippet)

Batch CSV / Multi-Line Evaluator (/v1/batch) Gear 1 Bulk

One State per Line (Up to 500 items)

Results, Hybrid Comparator & Token Attribution Ready

Click "Evaluate Decisions" to inspect calibrated distributions, conformal prediction sets, token attributions, and option-order invariance.

OpenJev-RLCD & Calibration Lab (Stage 1 Pathwise + Stage 2 LOO-KL) arXiv:2609.38850 & 2610.17309

Click "Run RLCD² Lab" to execute Algorithm 1 locally and visualize Brier Score reduction, ECE improvement, and the certified Clopper-Pearson threshold.

540-Item Reproducible Benchmark Suite (/v1/benchmark)

Loading 540-item evaluation corpus metrics...