Agent interface
Controls user interaction and workload input.
User-facing AI experiences that execute through Tenzarch.
AI Agents translate user intent into workloads. They are one source of execution demand, while the network behind them handles state, orchestration, routing, distributed processing, and results.
An Agent is an experience layer over the execution network.
The Agent defines how a user provides intent and receives output. It does not replace Execution Jobs, Workflow Runs, routing, or Execution Nodes. Those network responsibilities remain consistent regardless of which Agent originated the workload.
01User Intent
02AI Agent
03Execution Job
04Workflow Run
05Routing
06Execution Node
07Result
08Agent Experience
Keeping Agent UX separate from execution makes the platform extensible.
Execution state
Tracks the workload independently of the interface.
Infrastructure
Processes assigned work without becoming the Agent.
Result delivery
Returns output while preserving execution history.