The technical pipeline of how our engine ingests, cleans, and validates metadata for streaming platform submission — from raw payload to certified delivery, in under 0.3 seconds.
Metadata is received via API (JSON), file upload (DDEX XML, CSV, XLSX), or direct integration with a distributor's pipeline. Every payload is immediately hashed (SHA-256) and stored in immutable, encrypted storage. The original file is never modified — all subsequent processing operates on an in-memory copy.
The ingestion parser converts the incoming format into our internal canonical schema — a standardized JSON structure that represents metadata in a format-agnostic way. This normalization step ensures that regardless of how metadata arrives, it's processed identically downstream. Character encoding is normalized to UTF-8, whitespace is trimmed, and empty fields are flagged.
The core engine validates every field against 2,400+ rules organized by target store specification. Each field is checked for type, format, length, character encoding, and store-specific requirements. Validations run in parallel across all target stores — a release destined for 6 stores is validated against all 6 simultaneously, not sequentially.
The engine cross-references identifiers against external registries — MusicBrainz for recording data, ISRC registries for code validity, and store-specific taxonomies for genre mapping. This enrichment step catches mismatches that field-level validation alone cannot detect: an ISRC that's structurally valid but assigned to a different recording, or a genre tag that maps to the wrong browse category.
Validation results are compiled into a field-by-field compliance report. Each issue includes the field name, the store that flagged it, the severity (error or warning), the rule that was violated, and a suggested fix. The report is structured as JSON for API consumers and rendered as a PDF for human review. Every result is timestamped and bound to the original payload hash.
If validation passes (or passes with warnings), a compliance certificate is generated with a SHA-256 hash binding the metadata payload to its validation result. The certificate is stored immutably for 7 years. The validated metadata and certificate are then ready for delivery to the target stores via the distributor's standard delivery pipeline — or via our API for automated delivery.
Run your metadata through the full pipeline and get a field-by-field compliance report in under a second.