Skip to main content
This guide explains document processing from an integration perspective. It focuses on externally visible behavior and API results.

Processing lifecycle

When a document enters Nectar, it follows this lifecycle:
  1. Ingestion — The document is received from a connection, upload, or API submission.
  2. Extraction — Core fields (dates, totals, usage, meters, line items) are captured.
  3. Validation — Consistency checks are applied before data is published.
  4. Publication — Structured records become available in the API and platform.

What Nectar extracts

Nectar processes each document into structured entities you can query:
CategoryTypical outputs
Bill/document metadataBill date, period dates, invoice ID, total charges
Usage dataConsumption values, units, date ranges
Meter contextMeter identifiers and site/account associations
Line itemsCharge rows with descriptions, amounts, and categories
Demand data (when present)Demand usage and demand-related charges

Validation behavior

Before publishing data, Nectar applies quality checks such as:
  • Date consistency (billing period ordering)
  • Charge consistency (line-item totals vs. bill totals)
  • Duplicate detection (same document processed more than once)
  • Unit consistency (usage units aligned to expected commodity usage)
If a document needs attention, it may be flagged for review in Data Quality.

Multi-meter and complex documents

For documents that cover multiple meters, Nectar associates usage and charges to the best available meter context from document evidence. If a document includes account-level charges that are not meter-specific, those charges are represented at the appropriate aggregate level. For specialized utility formats (for example, demand electricity or waste services), Nectar preserves the source intent while keeping output fields consistent with the API schema.

What to expect in the API

Your integration should treat Nectar responses as the source of truth for:
  • Documents (processed bills)
  • Usage data (meter-level or account-level consumption records)
  • Line items (detailed cost components)
  • Connections and statuses (collection state and error handling)
For best results, integrate against v2.2 endpoints for all new builds.

Next steps