Cross-source identifiers
Every current identifier attachment, with the link method and confidence carried verbatim — deterministic links are 1.000; probabilistic links surface their real matcher confidence, never rounded up.
| Scheme | Value | Source table | Link method | Confidence | Since | Provenance |
|---|---|---|---|---|---|---|
| cms_op_manufacturer_id | 100000136433 | cms_open_payments | deterministic_manufacturer_id | 1.000 | 2026-06-13 | evidencerecorded: 2026-06-13 methodology: entity-graph/v1 evidence: {"rule":"D-OP-MANUFACTURER","manufacturer_id":"100000136433","manufacturer_name":"SI-TECHNOLOGY, LLC","manufacturer_state":"CO"} |
| fda_firm_name | SI-TECHNOLOGY LLC | fda_firm_link | deterministic_firm_name | 1.000 |
Graph edges
One-hop relationships — ownership, chain membership, operation — each backed by a federal source row.
| Relationship | Entity | Confidence | Source table | Provenance |
|---|---|---|---|---|
| pays → | EDWARD DONNER (CO) | 1.000 | cms_open_payments | evidence |
Current record
Every current claim asserted about this entity, grouped by provenance subject. Each cell carries its upstream federal source.
Point-in-time
Reconstruct this record as of any date, or diff it between two dates — both axes of the bitemporal store (bitemporal-read/v1): valid time (when a fact was true per the source) and recorded time (what the store knew when).
Pick a date to reconstruct, or two dates to diff. Equivalent JSON: /api/v1/history/{subject_type}/{subject_id}?at=… and …/diff?from=…&to=…