Anvia

Team runs ​

Studio can serve registered Agent teams over HTTP with attributed JSONL event streams, coordinator steering, cancellation, and application-owned interaction responses. The same runs are visible in the browser console; see Run a team.

Register a team ​

Pass team definitions alongside agents and pipelines. Agent and team IDs occupy separate namespaces, and members are not automatically registered as standalone agents.

ts
import { Agent, AgentTeam } from '@anvia/core/agent'
import { Studio } from '@anvia/studio'

const researcher = new Agent({ id: 'researcher', model })
const team = new AgentTeam({ id: 'research-team', model, members: [researcher] })

await new Studio([researcher, team]).serve({ port: 4021 })

Duplicate team IDs are rejected. GET /teams and /config list team IDs, registered member definitions, and limits; GET /teams/:teamId returns one definition.

Run and control a team ​

EndpointRequestBehavior
POST /teams/:teamId/runs{ prompt } or { messages }Start an attributed JSONL event stream
POST /teams/:teamId/runs/:runId/steer{ prompt } or user-only { messages }Queue user input for the coordinator
POST /teams/:teamId/runs/:runId/cancel{}Cancel the entire team
POST /teams/:teamId/runs/:runId/interactions/:interactionIdAgentInteractionResponseResolve one pending approval or question

The stream begins with { type: 'team_run_started', teamId, runId }; the same Studio control ID arrives in the x-anvia-team-run-id response header. Subsequent events follow the core AgentTeamEvent union — with core's teamRunId, instance IDs, parent IDs, and depth — followed by a terminal outcome or an error event. Internal agent continuation events are omitted; attributed interaction events carry everything an approval or question UI needs. StudioTeamRunRequest, StudioTeamRunEvent, and StudioTeamConfig are exported for clients.

Team requests accept at most 1 MiB of JSON and 256 messages; oversized bodies return 413.

Resolve interactions ​

Runs and pending interactions are scoped to the team and Studio run ID. Several members can request approval simultaneously, and only the HTTP application can answer them: an invalid reply leaves the interaction pending, and repeated replies are rejected. Apply authentication and authorization middleware to these routes when exposing Studio remotely, as with other Studio execution routes.

Lifecycle boundaries ​

  • Disconnecting the stream or shutting down Studio cancels the team and its pending interactions.
  • Completed runs leave the live registry; later control requests return 404.
  • Core's configured event buffer limit (maxBufferedEvents) bounds unread events.
  • Team runs use ephemeral conversations and are not stored in Studio's agent session history; the API provides no reconnection or durable resume.

Built for Anvia.