Orkena
▸ Browse docs — Core concepts

▸ SECTION B · Core concepts

Nodes

Last updated: 2026-08-16

A node is one step in a graph. The 19 node types:

Execution

NodePurpose
agentA model-backed agent step. Config: model, system prompt, temperature, tools.
toolAn MCP tool invocation. Every call is a policy decision point.
transformPure data transformation between steps.
routerConditional branch on a field or classifier output.
terminateEnds the run.
triggerStarts the run: inbound email, webhook, schedule, event.

Gates and human work

NodePurpose
gate.hitlHuman approval with quorum, SLA, escalation.
gate.policyPolicy evaluation gate (in addition to edge-level policy).
gate.evalEvaluation-suite gate — blocks until the suite passes.
human_taskAn assignment to a human (not an approval — work).

Structure

NodePurpose
mapFan-out over a list; each item runs the subgraph.
subgraphNested graph reference.
waitPause for a duration or a condition.

Memory

NodePurpose
memory.readRead from the workspace memory store.
memory.writeWrite to the workspace memory store (a ledger event).
rag.retrieveRetrieval from a connected vector store.

Example config

nodes:
  triage:
    type: agent
    model: anthropic/claude-sonnet
    tools: [search, classify]
  approve:
    type: gate.hitl
    quorum: 2
    roles: [ops_lead, compliance]
    sla: 4h
    escalate: vp_ops

Every node type is covered in the DSL reference with its full schema.

Was this helpful?