/sim/runsadminStart a simulation run: a population drawn from a pack (or a selection of personas), a stimulus, measures, a voiced tier (free text rated by SSR, with collapse checks) and dynamics (adoption over mixing groups, or opinion over an explicit graph). Runs as `simulate-voice` then `simulate` jobs; watch sim.progress / sim.done in /events. At most 3 queued or running per tenant (429).
Body
pack | the pack whose dimensions the population is drawn from |
population | {simulate: rows, represents?: people (each row stands for represents/simulate), seed?, segments?: [{name, share, values: {dimension: value | {shares} | {range: [lo, hi]} | {distribution}}}], rake?: [category dims] | {dim: {label: share}}, mixing?: {by: [dims or a mixing_group edge], edges?: {dim: [cut points]}}} or {personas: {where?, limit?, handles?, copies?, years_before? (a self-linked persona begins that many years younger, to live its past up to its age now)}, represents?, mixing?} |
stimulus | {text, kind?, structure?, images?: [at most 4 base64 data: URLs (jpeg, png, webp), each under 400,000 characters], shown to every voiced member with the words} |
measures | [purchase_intent | appeal | share_intent | attend_intent | {key, anchors: [5 statements low..high] | [[..], [..]], ask?}] (default purchase_intent) |
voice | {members: 0..500 voiced by the tenant's model, rated by SSR; retest?: 0..20 (2); prior?: true; expected_sd?: 1.1; temperature?; strata?: [category dims or "segment"] for a weighted stratum sample} |
dynamics | {model: adoption, channels: {name: rate/day | {default, general, <segment>, "dim=value", "scalar<25", "scalar>=65", "scalar=18..24": rate} (every key is checked: an unknown one is refused by name)}, imitation?, across?: 0..1, fatigue?, fatigue_shape?, appeal?: scalar dim, appeal_measure?, events?: [{at, exposure, half_life?, duration?, who?}], event_process?: {rate: {mean, per}, exposure, half_life?, who?}} | {model: opinion, graph?: {edges: [[listener, speaker, w?]]}, homophily?: [dims], stubbornness?: 0..1, rate?} |
step | any duration ("10min", "1h", "1d") |
horizon | a duration ("40d") |
report_every | a duration |
chunk | rows per chunk |
seeds | 1..20 (spread across seeds) |
reference_date | YYYY-MM-DD the run's time starts at (default today): durations map onto the calendar from it, and derived dimensions (a cohort from age) count years from it |
hold | true: created but not started until POST /sim/runs/:id/observations {release: true} |
observe | {tool, args?, members?: voiced|all, max?: 1..5000}: pull per-member observations from one of the tenant's read-only tools first |
calibration | {targets: [{measure?, metric: mean|would_act|have_it_end|ever_had_it_end, target, tolerance?}]}: checked in results.calibration |
