EPA Stationary Combustion · Inventory Completeness Assessment
Decides whether a declared emission source belongs in a stationary combustion inventory, counts as a completeness gap, or falls under other guidance.
Overview
For each declared emission source, the package decides whether it is accounted for in the stationary combustion inventory, represents a completeness gap that needs closing, or belongs to a different emissions category entirely, and grades the quality of the fuel-use data behind it — driving corrections to activity data and gap-closure work ahead of inventory finalization.
The scope follows EPA stationary combustion inventory guidance, covering boilers, furnaces, internal combustion engines, turbines, flares, process heaters, incinerators, and cooling systems; sources declared as mobile combustion, purchased electricity, air conditioning/refrigerant equipment, or process and fugitive emissions are routed out to their own dedicated guidance rather than assessed here.
In practice, a boiler dropped from the inventory as immaterial before it was ever assessed comes back as a gap, even when it also lacks activity data — assessment has to happen before materiality can be claimed. A mobile combustion source never counts as a stationary gap at all; it is routed out of scope outright. Floor-area-prorated fuel estimates for a shared office boiler are accepted as legitimate estimated data rather than flagged as missing. A metered turbine sharing fuel with vehicles still gets accounted for, but carries a double-counting warning and a flag for an outstanding regulatory cross-check.
When a source triggers both an improper exclusion and a genuine data gap, the exclusion is the reported reason, since the guidance treats assessment as a precondition for any materiality call — the missing data alone would not have disqualified the source.
Every run returns a structured verdict — accounted, gap, or out of scope — plus an activity-data-quality grade and any completeness flags with citations to the underlying guidance, ready to feed directly into inventory review or audit workflows.
What it decides
Try it — live sandbox
Build a case and run it against the publisher's live production release — the exact rules a subscriber calls. Run the same case twice and the answer is identical: no LLM in the loop.
Run a case to see the deterministic decision — tags, routes, reasons, evidence, and facts. Same case in, same answer out, every time.
Integrate — two ways
Call it directly over HTTP, or connect your coding agent over MCP. Same decision, same metering, either way.
One authenticated POST — the same deterministic, metered evaluation the sandbox runs, no LLM in the loop.
# Request shape only. Replace sert_mk_YOUR_KEY and send every field from the Inputs contract above.
curl -X POST https://api.sertainly.ai/api/marketplace/v1/packages/stationary_combustion_inventory_completeness_assessment/evaluate \
-H "Authorization: Bearer sert_mk_YOUR_KEY" \
-H "content-type: application/json" \
-d '{ "case": { "reporting_period_start": "example", "reporting_period_end": "example", "facility_identifier": "example" } }'