Dataset: AI Usage
Entity: AI Usage Measurement
Field ID: id
Type: Text
Description: The ID assigned to the record by the source application. AI providers publish daily counts rather than identified records, so this ID is generated by the Keypup connector: it is a short provider namespace followed by a digest of everything that distinguishes the measurement - the day, the aggregation level, the breakdown, the subject, the surface, the model and the client where applicable.
Source: Calculated
Transformation logic: The identifying components are hashed rather than written out, because record IDs are visible to end users and the components include personal identifiers such as email addresses and provider user IDs.
App Mapping
GitHub Copilot | Calculated: |
Anthropic Claude (coming soon) | Calculated: |
Cursor (coming soon) | Calculated: |
OpenAI Platform (coming soon) | Calculated: |
OpenAI Codex (coming soon) | Calculated: |
Reporting Use Cases
The id field is stable across re-reads, so a measurement that a provider restates keeps the same ID and is updated in place rather than duplicated.
Support and troubleshooting: The namespace prefix makes it immediately clear which integration produced a record, which is what makes this field worth showing in a drilldown.
Do not parse it: The digest is not readable and is not meant to be. Everything that went into it - day, aggregation level, breakdown, actor, surface, model, client - is separately available as a field, so no report needs to take the ID apart.