| Abstain | The result had no answer. The response’s status is abstained; an outcome aggregation’s abstain names the outcome to use, and may reuse one such as escalate. | From answers to an outcome |
| Acceptance | A suggestion is accepted when the version it proposed first serves primary production traffic after the suggestion was made. Previews, test keys and reverts never count. | Suggestions and previews |
| Aggregation | A feature Moonbounce computes from other features, such as a weighted_mean, a max or bands. No engine is involved, and nothing is learned. | Aggregations |
| Answer | What the engine returns for a question. A noul question’s answer is the probability that it is true, from 0 to 1. | From answers to an outcome |
| Attribute | A feature your app supplies with each call, such as a region or your own classifier’s score, declared in the structure and sent in attribute_values. The engine never sees it. | Describe a decision |
| Binding | Which engine a decision runs on, per environment. Changing it appends a new binding revision. | Engines, keys and environments |
| Check | An all or any aggregation: it compares inputs to thresholds and outputs 1 or 0. | Aggregations |
| Claim | One statement inside a feedback submission. | Feedback |
| Cut point | A value where a bands aggregation moves from one outcome to the next. In the refund decision: below 0.38 deny, from 0.38 escalate, from 0.72 grant. Part of the structure hash. | From answers to an outcome |
| Decision | A decision id you choose, such as refund-exception, plus its structure. | Describe a decision |
| Decision file | The file in your repository, by convention moonbounce/<decision_id>.json, holding decision_id and structure. | Versions and hashes |
| Dry run | moonbounce.dry_run: true on evaluate: runs the engines, stores no evaluation, and returns a dry_ id that names nothing stored; only its spend is recorded. | Evaluate content |
| Engine | The model service that answers a decision’s questions. Two are offered, Jev (the default) and Clef. | Engines, keys and environments |
| Engine profile | An opaque identity for the engines and exact models that answered an evaluation. Compare it for equality; use engines for the details. | Engines, keys and environments |
| Environment | Production or test. It comes from the API key, and each environment has its own evaluations, feedback, review queue and engine binding. | Engines, keys and environments |
| Evaluation | One run of a decision on one item: every answer, every aggregation’s result, the status and outcome, and the version that produced them. Its id starts evl_. | Evaluate content |
| Evidence label | How strong a preview’s evidence is, by what it was tested on: Starting examples, Training cases or Held out. | Suggestions and previews |
| Feature | A named value: a question’s answer, an attribute’s value, or an aggregation’s result. Questions, attributes and aggregations share one set of names. | Describe a decision |
| Feedback | What was true about an evaluation, sent by you, your app or a reviewer. It’s what previews and suggestions learn from. | Feedback |
| Idempotency key | A value you send in the Idempotency-Key header so a retried request returns the first result instead of running twice. | Evaluate content |
| Inputs | The features an aggregation reads, listed by path in its inputs. bands and match name a single input. | Aggregations |
| Item | The thing being judged, named by moonbounce.item_id, such as a ticket. Several evaluations of the same item share its id. | Evaluate content |
| Outcome | The named action your app takes, such as grant, produced by a bands, rules or match aggregation. Your code decides what it means. | From answers to an outcome |
| Owner | The person whose judgment decides a case. Feedback with source: "owner" comes from them. | Feedback |
| Path | How a structure refers to something: questions.<name>, attributes.<name> or aggregations.<name> for a feature, state.<key> (or deeper, state.order.total) for part of the state. | Describe a decision |
| Preview | A replay of a changed decision on your past cases with feedback, before you merge it: what it fixes, what it breaks, and an evidence label. | Suggestions and previews |
| Primary traffic | Evaluations your app acts on, as opposed to seed examples and replays. Only primary production traffic counts toward acceptance. | Evaluate content |
| Question | A feature the engine answers from the state. System One has four kinds; noul (yes/no) is the one available today. | Describe a decision |
| Reads | The parts of the state a question sees, listed in its reads as paths: a top-level key (state.order) or a key inside an object (state.order.total). The question gets just those values, nested as in the state. Without it, the question sees the whole state. | Evaluate content |
| Result | The final feature, named by the structure’s result as a path. It produces a value or an outcome. | Describe a decision |
| Review | Answering an item from the review queue in the web app. The person reviewing is a reviewer. | Review |
| Review queue | A short daily list of the evaluations whose review would teach the most, plus random spot checks. | Review |
| Scope | What an API key may do: decisions:evaluate, decisions:feedback or decisions:develop. | Engines, keys and environments |
| Seed examples | About a dozen labeled examples you upload when you start, so previews have cases to test against. They never count as traffic. | Feedback |
| Source | Who is speaking in a feedback submission: owner, app_native (your app’s own approve or reject feedback) or app_resolution (how your process finally settled the case). | Feedback |
| Spot check | A random item in the review queue, shown blind (without Moonbounce’s answer), so agreement can be measured honestly. | Review |
| State | The content the questions are asked about: any JSON value, including a string or JSON null. With an object, each question can read only some of its keys. | Evaluate content |
| Status | What kind of result an evaluation has: computed (a number, in value), decided (an outcome) or abstained (no answer). | From answers to an outcome |
| Structure | Everything that defines a decision: its questions, attribute declarations, aggregations and result. | Describe a decision |
| Structure hash | The hash of a structure, and so the decision’s version. Formatting and key order don’t change it; changing a question, a weight or a cut point does. | Versions and hashes |
| Suggestion | One proposed change to a decision, such as a reworded question or a new one, delivered as a diff with a preview. | Suggestions and previews |
| Threshold | The value a check or rule compares a feature against. Defaults to ≥ 0.5 on 0-to-1 values; any other threshold is written in the structure. | Aggregations |
| Value | The number a feature or the result produces. A question’s answer, a weighted average and a probability all run from 0 to 1. | From answers to an outcome |
| Version | A structure hash. There are no version numbers. | Versions and hashes |
