Skip to content
HickmaTechMLAtlas
Sign inRequest access
API docs preview

Clean integration previews for audit creation, status checks, and evidence downloads.

This page shows example requests against the live audit contract and the current download model, without claiming a complete developer portal yet.

Preview

API workflow preview

The API is designed for paid platform customers who want automated audit submission, run tracking, and evidence retrieval across Segment, Pattern, and Intersection dimensions.

The current backend uses the X-API-Key header. API-key-only examples are local development previews routed to admin@hickmatech.local; staging workspace requests are scoped by the signed-in session owner through the frontend proxy.
Dataset uploads, status polling, and artifact downloads follow the live audit contract, even though this page is still a compact preview rather than a full docs portal.
MLAtlas maps score behavior across Segments, Patterns, and Intersections to reveal robustness and reliability risks that global metrics can miss. Future integration points are documented in the README so this preview can evolve into real docs without redesigning the page.
Glossary

Key report terms

Plain-language definitions for the report terms that most often cause confusion during client review.

Classification threshold

The score cutoff used to turn predicted scores into positive or negative class labels. A threshold of 50.0% means scores at or above 50.0% are treated as predicted positives.

Global evaluation usability

Whether the overall audit produced enough valid evidence to interpret the top-level results. Usable means the aggregate evaluation can be relied on for governance review. Limited means a blocking warning reduced confidence in the overall evaluation.

Equivocal score band

The low-confidence score range configured for the audit. Scores inside this band are counted as equivocal rather than clearly positive or clearly negative.

High-risk and medium-risk gap thresholds

The percentage-point cutoffs used to label slice-level findings as HIGH or MEDIUM risk when compared with global performance.

Create audit draft

Create an organization-scoped score-driven audit draft.

bash
curl -X POST http://localhost:8001/orgs/$ORG_ID/audits \
  -H "X-API-Key: $MLATLAS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"display_name":"Synthetic score-driven audit"}'

Upload scored CSV

Upload a CSV or CSV.GZ under the organization and audit run.

bash
curl -X POST http://localhost:8001/orgs/$ORG_ID/audits/$AUDIT_ID/upload \
  -H "X-API-Key: $MLATLAS_API_KEY" \
  -F "dataset=@demo_data/synthetic_testing_scores.csv;type=text/csv"

Run audit

Run the configured v2 score-driven audit and generate official artifacts.

bash
curl -X POST http://localhost:8001/orgs/$ORG_ID/audits/$AUDIT_ID/run \
  -H "X-API-Key: $MLATLAS_API_KEY"

Download artifact

Download generated evidence through the org-scoped artifact route.

bash
curl -L http://localhost:8001/orgs/$ORG_ID/audits/$AUDIT_ID/artifacts/audit_summary.json \
  -H "X-API-Key: $MLATLAS_API_KEY" \
  -o audit_summary.json