Online conformance: replay an ORDERED event stream against a stored model and get a running verdict — does the model still match, or which event was the first that did not, and why. Unlike sim_conformance (whole finished cases, aggregate fitness) the order of events IS the stream's order, cases may interleave, and the answer names the first non-conforming event: its position, case, activity, whether the activity is unknown to the model or the transition was not enabled, the places that blocked it (need vs have), the case's marking just before it, and `repair` — the shortest unobserved transitions that would have enabled it (a local alignment of that one step, bounded; absent when none is found within repairDepth). This is token replay, not trace alignment. A divergent event is force-fired so the case resynchronises and later events are judged on their own. fitness is conforming events / events; tokenFitness is 1 - missing/consumed tokens: 1 while nothing has been consumed, and below 0 when more tokens are missing than were consumed. Missing tokens are counted as sim_conformance's replay counts them (an unknown activity and a blocking inhibitor each count one, an unmet read arc its shortfall), except where that replay ignores capacity or merges parallel arcs, which its caveats name. Give the stream as `events` (JSON array of {case, activity, timestamp?}) or `log` (CSV with a case_id, activity header, timestamp optional; rows are taken in the order given, not sorted). Activities are transition ids. Each case replays from the model's initial marking, so the model should be the per-case workflow; guard expressions are not evaluated and timestamps are not checked against delays (caveats says so). keepSteps adds the per-event running signal. Costs 2 tokens from your account's compute bucket. With outlook (JSON {"goal": {place: count}, "avoid": {place: count}, "cases": [case ids]}) the reply adds outlook, separate from conforming and firstDivergence: from the marking each case is in after the replay, whether goal is reached on every way the net can go on and avoid on none. Each target constrains only the places it names, to exactly that count, as sim_reach's to does. Each verdict is guaranteed, forced-violated, open (some continuations satisfy it and some do not) or unknown, which names the budget, cap, guard or timer that stopped the search and is never a yes or a no. A verdict rests on the case's marking itself, a sim_reach answer from that marking (reach; re-check a yes or an invariant no with sim_check_witness(reach=…)), an escape (a deadlock or a cycle that never reaches the target, re-checked before it is returned), or a search of every marking before the target, which has no smaller certificate. No fairness is assumed: a loop that can repeat forever makes a goal open, unless a delayed transition stays enabled along it unfired (the simulator starts a timer the instant it is enabled), which is unknown. Guards decidable from the marking are enforced, though the replay ignores them; a transition whose guard reads action parameters is never fired. Cases are grouped by marking, at most 4 distinct markings per call (outlook.cases, default every case in the stream, narrows it), and each search is bounded by sim_reach's 100,000-state and 20,000,000-unit budgets. The outlook costs 2 more tokens per distinct marking and predicate the marking does not already decide (one for each reachability search it may run, as sim_reach's search costs 1; at most 16), charged before any runs, even when the first search decides it; a refused outlook costs nothing more.