semiont

Compose a synthesized aggregate resource — walk many annotations bound to or about a single anchor, assemble markdown, yield a Resource whose purpose is to be read (not referred to)

Run the knowledge enrichment pipeline on a resource using @semiont/sdk — detect entity references, resolve them against the KB, and generate new resources for unresolved ones

Build a job-claim worker daemon — claim jobs from the queue, process them, and emit lifecycle events. Cross-package wiring with @semiont/sdk + @semiont/jobs + @semiont/http-transport + @semiont/observability.

Apply structural-analysis tag schemas to a Semiont resource — classify passages by their structural role using IRAC, IMRAD, Toulmin, or any KB-registered schema via mark.assist with motivation tagging

Add assessment annotations to a Semiont resource — flag scheduling risks, dangers, inaccuracies, logical gaps, or other evaluative concerns using AI-assisted or manual assessment

Help users accomplish knowledge work tasks using the Semiont CLI

Add commenting annotations to a Semiont resource — suggest edits, ask questions of the author, or point things out to readers using AI-assisted or manual commenting

Add highlighting annotations to a Semiont resource — mark key passages, important claims, or noteworthy content using AI-assisted or manual highlighting

Bootstrap a Semiont knowledge base from a corpus of source files — declare the entity-type vocabulary via frame, then upload one resource per file via yield.resource

Extract relationships between canonical nodes — run a second mark.assist linking pass after the node set exists, with relationship-vocabulary instructions, so the resulting annotations link two nodes apiece

Drive @semiont/sdk's SemiontSession for long-running scripts — token refresh, bus event subscription, lifecycle observables, graceful shutdown

Install, provision, and run Semiont locally using published npm packages — no repo clone needed