Skip to main content

​
useAgentSession

Within applications deploying embedded AI components or requiring stringent access validation prior to execution, the useAgentSession hook operates as the primary initialization interface. It handles secure one-time ticket exchanges, retrieves the current session state, and dictates the active conversational agent for the application instance. By interfacing with the /agent/session API endpoint, the hook ensures that subsequent operations executed via useAgentContext are fully authenticated and explicitly routed.

​
Configuration Parameters

​
ticket
string | null
An optional, single-use authentication ticket mapped to a backend grant. When supplied, the hook exchanges this token sequentially before fetching session parameters. Omit or pass null if navigating a standard, persistent session.

​
Hook Interface

​
hasSession
boolean
Indicates the verified existence of an active and valid agent session.
​
sessionLoading
boolean
Indicates whether the hook is actively resolving the required session state via network requests.
​
sessionError
Error | null
Captures errors resulting from malformed tickets, expired sessions, or network failures.
​
sessionId
string | null
The unique identifier mapped to the active agent session.
​
contextGroup
string | null
An organizational assignment key applied specifically to this session structure.
​
agents
AgentWithIntegrations[]
A serialized array of AI Agents explicitly available within the verified user’s authorization scope.
​
activeAgent
AgentWithIntegrations | null
The specific agent currently designated for active interaction processing.

​
Methods

​
setActiveAgent
(agent: AgentWithIntegrations) => void
​
ticketExchanged
boolean
A state indicator confirming that a strictly provided authorization ticket has successfully exchanged for a session footprint.
​
ticketLoading
boolean
Indicates whether the specific ticket exchange operation is actively processing.
​
refetch
() => Promise<void>

​
Implementation Guidelines

​
Dynamic Agent Selection

In multi-agent environments, explicit selection interfaces are required to ensure the user communicates against the proper context model: 
import { useAgentSession } from "@wacht/tanstack-router";

function TargetAgentSelector() {
  const { sessionLoading, agents, activeAgent, setActiveAgent } = useAgentSession();

  if (sessionLoading) return <div className="text-gray-500 font-medium">Validating agent configurations...</div>;
  if (agents.length === 0) return <div className="text-red-700">No agents authorized for current session parameters.</div>;

  return (
    <div className="flex flex-col gap-3">
      <label htmlFor="agent-selector" className="text-sm font-semibold uppercase tracking-wide text-gray-700">
        Designate Operating Agent
      </label>
      <select
        id="agent-selector"
        className="form-select block w-full pl-3 pr-10 py-2.5 text-base border-gray-300 focus:outline-none focus:ring-blue-500 focus:border-blue-500 sm:text-sm rounded-md shadow-sm"
        value={activeAgent?.id ?? ""}
        onChange={(e) => {
          const selectedTarget = agents.find((agent) => agent.id === e.target.value);
          if (selectedTarget) setActiveAgent(selectedTarget);
        }}
      >
        <option value="" disabled>Awaiting agent selection...</option>
        {agents.map((agent) => (
          <option key={agent.id} value={agent.id}>
            {agent.name}
          </option>
        ))}
      </select>
    </div>
  );
}

​
Initializing Embedded Security Contexts

When configuring Wacht embedded contexts utilizing backend-generated session tickets, the hook must process the one-time artifact before rendering the interface: 
import { useAgentSession } from "@wacht/tanstack-router";

function SecureEmbeddedContext({ authorizationTicket }: { authorizationTicket: string }) {
  const { ticketLoading, hasSession, sessionError } = useAgentSession(authorizationTicket);

  if (ticketLoading) {
    return (
      <div className="flex animate-pulse items-center justify-center p-12 text-indigo-700 font-semibold tracking-wide">
        Exchanging authorization token...
      </div>
    );
  }

  if (sessionError) {
    return (
      <div className="p-6 bg-red-50 border border-red-200 rounded-lg">
        <h3 className="text-red-800 font-bold mb-2">Authentication Subsystem Error</h3>
        <p className="text-red-600 font-mono text-sm">{sessionError.message}</p>
      </div>
    );
  }

  if (!hasSession) {
    return <div className="p-6 text-gray-500 italic text-center">Security context expired or rejected. Request new authorization parameters.</div>;
  }

  return (
    <div className="p-8 border border-gray-200 rounded-xl bg-white shadow-md">
      <h2 className="text-xl font-bold text-gray-900 border-b pb-4 mb-4">Secure Context Initialized</h2>
      <p className="text-gray-600 leading-relaxed">
        The application is now permitted to initiate execution streams against local operations.
      </p>
      {/* TargetAgentSelector and Chat components mount below */}
    </div>
  );
}
  • useAgentContext - Implemented upon agent selection to deploy the associated execution stream.
  • useAgentIntegrations - Review available integrations corresponding to the designated active agent.