Creates a cryptographically anchored attestation for a dataset, recording its provenance, integrity hash, and schema version. The attestation is linked to the originating agent and anchored in the transparency log for full auditability.
Dataset attestations enable downstream consumers to verify that training data, evaluation sets, or inference inputs have not been tampered with and can be traced back to a known source.
Authentication
API key with datasets:write scope. Alternatively, pass a Bearer JWT token in
the Authorization header.
Tenant identifier for multi-tenant isolation.
Request
MAIP agent identifier that produced or owns this dataset (e.g.
maip-agent:01HXYZ...).
Human-readable dataset name. Must be unique within the agent’s namespace.
Semantic version of the dataset (e.g. 1.0.0). Defaults to 1.0.0 if
omitted.
SHA-256 hex digest of the dataset contents. Used for integrity verification at
consumption time.
Total number of records in the dataset.
Version identifier for the dataset’s internal schema (e.g. v2.1).
URI pointing to the dataset’s storage location (e.g.
s3://bucket/path/data.parquet).
Arbitrary key-value metadata attached to the attestation. Useful for tagging,
categorization, or linking to external systems.
Response
Unique dataset attestation identifier in MAIP format (maip-ds:ULID).
The agent that attested this dataset.
Dataset name as provided in the request.
Resolved version of the dataset.
SHA-256 hex digest stored for integrity verification.
Attestation status. Always attested on creation.
Transparency-log receipt anchoring this attestation.
ISO 8601 timestamp of creation.
API key for machine-to-machine authentication
SHA-256 hash of the dataset
Data format (e.g. parquet, csv, jsonl)
Additional dataset metadata