SemanticHub documentation
From public sources to a structured result and safe delivery.
SemanticHub monitors sources, judges whether new material is relevant, groups similar articles into clusters, and hands the result to your newsroom or another system.
- 1Fetch
- 2Relevance
- 3Embedding
- 4Clustering
- 5Workflow
- 6Result
- 7Review
- 8Delivery
Start here
How SemanticHub worksFollow the path from crawl to delivery and learn the core concepts.Run your first goalSet up monitoring, collect a first cluster, and inspect the result without sending anything to production.Your first clusterHow to read a cluster, when to accept it, and when to reject it.
Run a topic
SourcesWhat the system fetches, from where, and how often.Workflows and resultsDirect mode, or processing through successive AI steps.Knowledge basesGive agents PubMed, Wikipedia, or your own API, and attach them to a goal.
Connect the system
WordPress pluginReceive finished texts in WordPress as drafts or published posts.WebhooksDelivery contract, retries, idempotency, and test mode.API tokensCreate and rotate secrets for supported integrations.