[{"data":1,"prerenderedAt":314},["ShallowReactive",2],{"docs-index":3},[4,104,184],{"id":5,"title":6,"body":7,"description":95,"extension":96,"meta":97,"navigation":98,"path":99,"seo":100,"stem":101,"updated":102,"__hash__":103},"docs\u002Fdocs\u002F1.getting-started.md","Getting started",{"type":8,"value":9,"toc":88},"minimark",[10,19,22,27,30,52,56,67,70,74],[11,12,13,14,18],"p",{},"TameChaos is an ",[15,16,17],"strong",{},"edge-first experimental control platform",": you design sensor-driven\nsystems, run them on the authoritative edge, and let a fleet of edges improve its models\nfrom its own operation. This guide orients you to the pieces and points at the rest of\nthe docs.",[11,20,21],{},"These pages are documentation for the marketing site — the product itself lives in the\ncloud app. Use them to understand the model before you open the app.",[23,24,26],"h2",{"id":25},"the-shape-of-the-platform","The shape of the platform",[11,28,29],{},"Everything you build follows the same three steps:",[31,32,33,40,46],"ul",{},[34,35,36,39],"li",{},[15,37,38],{},"Design"," — compose real and virtual sensor arrays, physics and inference models, and\nthe experiments that exercise them.",[34,41,42,45],{},[15,43,44],{},"Run"," — execute experiments and simulations on the edge, in real, simulated, or\nhybrid mode, deterministically and repeatably.",[34,47,48,51],{},[15,49,50],{},"Improve"," — aggregate like-processes across your fleet to refine the models that\ndrive your systems.",[23,53,55],{"id":54},"edge-first-cloud-second","Edge first, cloud second",[11,57,58,59,62,63,66],{},"The single most important idea is where execution lives. The ",[15,60,61],{},"edge server is the\nauthoritative runtime"," — it owns your hardware and runs the control loop without waiting\non the network. The ",[15,64,65],{},"cloud is a non-authoritative layer"," that oversees the fleet and\naggregates across it; it never starts, stops, or mutates a live run.",[11,68,69],{},"If you remember one thing, remember that split. The rest of the docs build on it.",[23,71,73],{"id":72},"where-to-next","Where to next",[31,75,76,82],{},[34,77,78,81],{},[15,79,80],{},"Core concepts"," — the vocabulary: sensors, experiments, runs, and process-class\ntemplates.",[34,83,84,87],{},[15,85,86],{},"Edge and cloud"," — exactly which responsibilities sit where, and why distribution is\npublish-and-pull.",{"title":89,"searchDepth":90,"depth":90,"links":91},"",2,[92,93,94],{"id":25,"depth":90,"text":26},{"id":54,"depth":90,"text":55},{"id":72,"depth":90,"text":73},"A short orientation to TameChaos — what the platform is, how the edge and cloud divide responsibility, and where to go next.","md",{},true,"\u002Fdocs\u002Fgetting-started",{"title":6,"description":95},"docs\u002F1.getting-started","2026-06-01","4UI8sTxvLJ0bOBs8DyAIhdfLvvu1uGc0j7VGVhmskJw",{"id":105,"title":80,"body":106,"description":178,"extension":96,"meta":179,"navigation":98,"path":180,"seo":181,"stem":182,"updated":102,"__hash__":183},"docs\u002Fdocs\u002F2.core-concepts.md",{"type":8,"value":107,"toc":172},[108,111,115,122,126,137,147,151,162,165,169],[11,109,110],{},"A handful of terms carry most of the platform. They compose cleanly: sensors feed\nexperiments, experiments produce runs, and process-class templates make runs comparable\nacross the fleet.",[23,112,114],{"id":113},"sensor-arrays","Sensor arrays",[11,116,117,118,121],{},"A ",[15,119,120],{},"sensor array"," is a set of inputs to a system — built from physical hardware, replay\nof recorded data, or model-backed virtual sensors. The abstraction is the same across all\nthree, so an experiment does not care whether a channel is a real thermocouple or a\nsimulated one.",[23,123,125],{"id":124},"experiments-and-runs","Experiments and runs",[11,127,128,129,132,133,136],{},"An ",[15,130,131],{},"experiment"," is the definition of something to exercise — the sensors, the models in\nthe loop, and the conditions to apply. A ",[15,134,135],{},"run"," is one execution of an experiment on the\nedge, in real, simulated, or hybrid mode. Runs are deterministic and replayable: given a\nseed and a fixed timestep, a simulated run reproduces.",[138,139,144],"pre",{"className":140,"code":142,"language":143,"meta":89},[141],"language-text","experiment (definition)  ──run──▶  run #1  (edge A, real)\n                          ──run──▶  run #2  (edge A, simulated)\n                          ──run──▶  run #3  (edge B, hybrid)\n","text",[145,146,142],"code",{"__ignoreMap":89},[23,148,150],{"id":149},"process-class-templates","Process-class templates",[11,152,117,153,156,157,161],{},[15,154,155],{},"process-class template"," describes a ",[158,159,160],"em",{},"kind"," of process, independent of any single\nedge. When runs across the fleet instantiate the same template, they become comparable —\nthat shared identity is the join key the cloud aggregates on.",[11,163,164],{},"This is what makes fleet-level learning possible: without a common template identity,\nruns on different edges are just unrelated time series. With it, they are a comparable\nset the cloud can aggregate and, in the model-loop phase, learn from.",[23,166,168],{"id":167},"how-they-fit-together","How they fit together",[11,170,171],{},"Sensors define what you measure; experiments define what you do with them; runs are the\nrecord of doing it; and templates are what let many runs — across many edges — line up\ninto something a fleet can improve against.",{"title":89,"searchDepth":90,"depth":90,"links":173},[174,175,176,177],{"id":113,"depth":90,"text":114},{"id":124,"depth":90,"text":125},{"id":149,"depth":90,"text":150},{"id":167,"depth":90,"text":168},"The vocabulary of TameChaos — sensor arrays, experiments, runs, and the process-class templates that make runs comparable across a fleet.",{},"\u002Fdocs\u002Fcore-concepts",{"title":80,"description":178},"docs\u002F2.core-concepts","4Uu7Gb64QoDkEKUsLN5sLsREz_gZqovOBWaYRUqJJ8Q",{"id":185,"title":86,"body":186,"description":308,"extension":96,"meta":309,"navigation":98,"path":310,"seo":311,"stem":312,"updated":102,"__hash__":313},"docs\u002Fdocs\u002F3.edge-and-cloud.md",{"type":8,"value":187,"toc":303},[188,199,203,206,210,217,281,285,296],[11,189,190,191,194,195,198],{},"TameChaos splits responsibility along one line: the edge is authoritative for\n",[15,192,193],{},"execution",", the cloud is authoritative for ",[15,196,197],{},"definitions and identity",". Getting this\nboundary right is what keeps real-world control correct under bad networks while still\nletting a fleet learn.",[23,200,202],{"id":201},"the-edge-is-authoritative-for-execution","The edge is authoritative for execution",[11,204,205],{},"The edge server decides which definitions it adopts and runs, and it owns the live runs\nthemselves. It executes the control loop locally, without waiting on the network. Nothing\nin the cloud can start, stop, or mutate a run — the edge holds that authority because it\nis the only place close enough to the process to be right about it.",[23,207,209],{"id":208},"the-cloud-is-authoritative-for-definitions","The cloud is authoritative for definitions",[11,211,212,213,216],{},"The cloud owns template and model ",[15,214,215],{},"definitions"," and their identity — the registry. It\nissues template identity, and it observes, aggregates, compares, and suggests. A template\nis an inert definition until an edge instantiates a run from it.",[218,219,220,236],"table",{},[221,222,223],"thead",{},[224,225,226,230,233],"tr",{},[227,228,229],"th",{},"Concern",[227,231,232],{},"Edge",[227,234,235],{},"Cloud",[237,238,239,251,261,271],"tbody",{},[224,240,241,245,248],{},[242,243,244],"td",{},"Live runs (start\u002Fstop\u002Fmutate)",[242,246,247],{},"Authoritative",[242,249,250],{},"Never",[224,252,253,256,258],{},[242,254,255],{},"Adoption of a template\u002Fmodel",[242,257,247],{},[242,259,260],{},"Suggests",[224,262,263,266,269],{},[242,264,265],{},"Template & model definitions",[242,267,268],{},"Pulls & runs",[242,270,247],{},[224,272,273,276,279],{},[242,274,275],{},"Fleet oversight & aggregation",[242,277,278],{},"Reports",[242,280,247],{},[23,282,284],{"id":283},"publish-and-pull-never-command","Publish-and-pull, never command",[11,286,287,288,291,292,295],{},"Because the edge is authoritative for adoption, improved templates and models are\n",[15,289,290],{},"published"," to the fleet and ",[15,293,294],{},"pulled"," by edges that choose to adopt them. The cloud\nnever pushes a change into a running system. This is the safe direction for real-world\ncontrol: the device that owns the process decides what it runs.",[11,297,298,299,302],{},"Model improvement and distribution are the ",[15,300,301],{},"model-loop phase, on the roadmap","; fleet\noversight and like-process aggregation are the v1 cloud surface today. The authority\nboundary above holds in both phases — the cloud's role grows, but it never crosses into\ncommanding the edge.",{"title":89,"searchDepth":90,"depth":90,"links":304},[305,306,307],{"id":201,"depth":90,"text":202},{"id":208,"depth":90,"text":209},{"id":283,"depth":90,"text":284},"Which responsibilities sit on the authoritative edge and which sit in the non-authoritative cloud — and why distribution is publish-and-pull, never command.",{},"\u002Fdocs\u002Fedge-and-cloud",{"title":86,"description":308},"docs\u002F3.edge-and-cloud","oXt630i6i8WLOwMhZvGbP5_ZdYoqKTYJk_uudgyBy6A",1780537671975]