Demo foundation

A typed runtime graph for agentic work.

The current demo shows the foundation: typed nodes, input/output contracts, runtime events, trace replay, payload inspection, and graph execution.

This is the foundation, not the finished platform.

Graph runner reimagined
Semantiv productized graph workbench with typed runtime graph, trace replay, node inspector, event stream, contracts, and evidence
01

Typed runtime graph

Nodes, edges, branches, and boundaries represent how work actually moves.

02

Contracts

Inputs, outputs, evidence, and final values become inspectable runtime objects.

03

Trace replay

Frames, events, values, and transitions can be reviewed after the run.

04

Payload inspection

Runtime values are tied back to the node and handoff that produced them.

Runtime typing matters when an AI output crosses a boundary.

The demo direction is not simply visualizing a workflow. The point is to ask whether the operation remains valid while it runs.

  • MissingWhat evidence is absent before the operation can continue?
  • MalformedWhere did output fail its expected shape?
  • ReplayWhich branch can be repaired without rerunning everything?
Boundary failure
Runtime boundary failure view showing contracts, evidence, missing approval, runtime value, and repair choices

Where the demo goes next.

The next layer turns traces and contracts into scenario tests, candidate operating structures, evidence requirements, and review paths.

Candidate comparison
Operation variant comparison surface with candidate structures and measured tradeoffs

Want to discuss the demo?

Book a consultation to see whether the typed runtime foundation maps to your product, workflow, or customer problem.