Technology
Voyager creates a searchable information layer by connecting to source systems, interpreting what it finds, enriching the resulting records, and indexing those records for discovery and use.
The basic processing flow is:
Source → connection and discovery → format identification and extraction → processing and enrichment → Search Index → user interfaces and APIs.
The steps
1. Connect
A Voyager connector defines how an Indexing Agent communicates with a configured source.
Sources can represent different kinds of systems and repositories. Exact connector behavior, authentication, change detection, and supported source versions depend on the connector and product release.
2. Discover and extract
An Indexing Agent discovers source objects and reads the information needed for indexing.
The Platform identifies formats and invokes extractors that can derive metadata, text, geometry, thumbnails, or other information from the source content.
3. Normalize and enrich
Processing pipelines can transform or enrich records before they are indexed.
Depending on the configured workflow, pipelines may map fields, normalize values, derive additional metadata, calculate spatial properties, classify information, invoke external resources, or perform other supported processing.
4. Index
The resulting records are written to Voyager’s Search Index. The index stores metadata, extracted content, and derived fields used for search, filtering, analytics, and retrieval. The authoritative source content generally remains under the source system's ownership.
5. Search and use
Voyager user interfaces and APIs can query the indexed catalog.
The catalog can support downstream workflows such as delivery, transformation, analytics, applications, or grounded AI experiences.
5. Keep information current
Voyager supports multiple approaches to maintaining indexed information as sources change, including scheduled and incremental processing and supported event-driven integration patterns.
Actual freshness depends on the source, connector, change-detection mechanism, configuration, processing workload, and infrastructure.
Voyager is designed to work with distributed authoritative sources without requiring all source content to be consolidated into a single repository. The exact metadata, derived information, or source content that is processed or moved depends on configuration.
start a conversation
