Glossary
Key terms used across HeadlessAnalytics documentation, API reference, and SDKs.
- Workspace
- The tenancy boundary for keys, events, entities, dashboards, connectors, decisions, proof trails, and exports. Every API key belongs to exactly one workspace.
- Business
- A graph node representing a company, brand, client, location, product line, or operating unit. Every workspace has a default business. Businesses group events, entities, and metrics.
- Event
- The atomic fact: something happened, caused by an actor, against an object, in a source/context, with properties and outcomes. Events are ingested via /api/ingest and stored in the operating graph.
- Trackable
- A generated or external identifier for QR codes, barcodes, RFID tags, sensors, robots, agents, MCP servers, and CLI workflows. Trackables link physical or digital identifiers to event streams.
- Proof Envelope
- Metadata returned on every query: queryId, workspaceId, businessId, period covered, generatedAt, and schemaVersion. Lets agents and humans verify data freshness and trace insight lineage.
- Operating Graph
- Events become entities, relationships, attribution touchpoints, agent runs, workflow runs, decisions, receipts, proofs, and outcomes. The graph is the central data structure.
- Full Hypervisor
- The governance layer for consequential actions: DecisionObject, AuthorityGrant, ContextContract, attestation evidence, EnforcementReceipt, and proof trail export.
- DecisionObject
- A proposed or approved consequential action (e.g., scale a campaign, release funds). States: draft, pending_approval, approved, escrowed, released, denied, revoked, expired.
- AuthorityGrant
- Delegates permission for a specific actor (agent, operator, system) to perform a specific governed action. Tied to a context contract.
- ContextContract
- Binds a governed action to a specific business, surface, and set of constraints. Ensures actions run within approved boundaries.
- EnforcementReceipt
- The result of a governance enforcement check: allow, deny, escrow, or needs_human_approval. Includes the receipt ID, action taken, and timestamp.
- Connector
- Normalized ingestion from enterprise systems: Salesforce, HubSpot, Shopify, Stripe, Google Ads, Meta Ads, GA4, QuickBooks, Xero, and more. Connectors pull or receive webhook data into the event graph.
- MCP
- Model Context Protocol — a standard for giving AI agents tools. HeadlessAnalytics exposes MCP tools for tracking, querying, recommending, enforcing, and exporting intelligence.
- GraphQL
- A flexible query surface for headless data access. Query metrics, funnels, business health, agent performance, and workflow status without REST endpoint proliferation.
- Trust Tier
- Classification of data provenance: T0 raw, T1 observed, T2 attributed, T3 authority-bound, T4 proof-ready. Higher tiers have more verification and governance.
- Schema Version
- Tracks the version of the event envelope format. Ensures agents and systems can handle format evolution without breaking.