DatoricQuality methodology
Quality you can
measure and audit.
Datoric turns each collection specification into modality-specific checks, review criteria, and delivery metadata your team can evaluate.
The pipeline
Quality is a process, not a promise.
Each program begins with a written specification and modality-specific review criteria. The exact checks vary with the data.
- 01
Specification-driven collection
Requirements are written down first, so contributors and reviewers are working against the same explicit definition of what good looks like.
- 02
Signal and metadata checks
The specification identifies the technical fields used for the modality, such as audio signal quality, video sync and visibility, or timestamped action metadata.
- 03
Redaction and safety review
Where personal or sensitive material is in scope, the specification defines the applicable review and redaction controls.
- 04
Labeling and consistency review
Label, transcript, and annotation review rules are set against the fields and acceptance criteria for the program.
- 05
Acceptance
Accepted material is evaluated against the agreed criteria, with the documented quality fields included in the applicable delivery metadata.
Primary evidence
The methodology is visible in what we publish.
Quality claims here are backed by the datasheets and research on this site.
| What we state | Primary evidence |
|---|---|
| Catalog specifications list the quality and review fields available for each dataset. | Dataset catalogChecked July 21, 2026 |
| The computer-use repository documents outcome review, PII review, and duplicate and malformed-trace filtering. | Computer-Use Traces repositoryChecked July 21, 2026 |
| We apply measured evaluation to models in published research notes. | Research notesChecked July 21, 2026 |
Questions
Quality methodology
- What does QA actually check?
- It depends on the modality. Catalog specifications identify available signal, sync, visibility, labeling, and review fields. The computer-use repository additionally documents human outcome review, PII review, and duplicate and malformed-trace filtering.
- Do you measure quality on your own benchmarks?
- Yes. Datoric Research publishes evaluation notes with public PDFs, methods, sample sizes, tables, and stated limitations.
- Can we set our own acceptance criteria?
- For build-to-spec programs, acceptance criteria are agreed with your team before collection begins, so QA is evaluating against the definition of done you care about.
Continue
More of the record
How it works
From spec to delivery: the end-to-end collection and review workflow.
Data rights & provenance
How consent, licensing, and origin are defined and reviewed.
Security & privacy
Request-based access, privacy handling, and program safeguards.
Licensing
Commercial license scope and the terms buyers review.
Custom data collection
Build-to-spec programs for data that does not exist yet.
About Datoric
Company identity, research group, and contacts.